Subsurface fluid storage framework

Abstract

A method can include controlling a simulator to perform a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, computing fugacity values using an equation of state, computing an equilibrium constant value using a first machine learning model, computing a Henry's law constant value using a second machine learning model, and computing component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry's law constant value; and generating simulation results for the subsurface region based at least in part on the component mole fraction values.

Description

SUBSURFACE FLUID STORAGE FRAMEWORKRELATED APPLICATIONS

[0001] This application claims priority to and the benefit of a U.S. Provisional Application having Serial No. 63 / 605,703, filed 4 December 2023, which is incorporated by reference herein.BACKGROUND

[0002] Various types of fluids may be stored in a subsurface region. For example, consider storage of hydrogen in a subsurface region, storage of carbon in a subsurface region, etc. Various field operations may include injection of fluid into a subsurface region and / or production of fluid from a subsurface region. Various technologies, techniques, etc., described herein pertain to operations that may involve storage of fluid and / or production of fluid.SUMMARY

[0003] A method can include controlling a simulator to perform a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, computing fugacity values using an equation of state, computing an equilibrium constant value using a first machine learning model, computing a Henry’s law constant value using a second machine learning model, and computing component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constant value; and generating simulation results for the subsurface region based at least in part on the component mole fraction values. A system can include a processor; a memory operatively coupled to the processor; processor-executable instructions stored in the memory and executable to instruct the system to: perform a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, compute fugacity values using an equation of state, compute an equilibrium constant value using a first machine learning model, compute a Henry’s law constant value using a second machine learning model, and compute component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constantvalue; and generate simulation results for the subsurface region based at least in part on the component mole fraction values. One or more computer-readable storage media can include processor-executable instructions executable by a system to instruct the system to: perform a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, compute fugacity values using an equation of state, compute an equilibrium constant value using a first machine learning model, compute a Henry’s law constant value using a second machine learning model, and compute component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constant value; and generate simulation results for the subsurface region based at least in part on the component mole fraction values. Various other apparatuses, systems, methods, etc., are also disclosed. This summary is provided to introduce a selection of concepts that are further described below in the detailed description. This summary is not intended to identify key or essential features of the claimed subject matter, nor is it intended to be used as an aid in limiting the scope of the claimed subject matter.BRIEF DESCRIPTION OF THE DRAWINGS

[0004] Features and advantages of the described implementations can be more readily understood by reference to the following description taken in conjunction with the accompanying drawings.

[0005] FIG. 1 illustrates an example system that includes various components for simulating a geological environment;

[0006] FIG. 2 illustrates an example of a system;

[0007] FIG. 3 illustrates an example of an architecture;

[0008] FIG. 4 illustrates an example of an equation of state;

[0009] FIG. 5A and FIG. 5B illustrates example plots;

[0010] FIG. 6 illustrates example plots as shown in detail in FIG. 7A to FIG.7L;

[0011] FIG. 7A to FIG. 7L illustrate example plots;

[0012] FIG. 8 illustrates an example plot;

[0013] FIG. 9A, FIG. 9B, and FIG. 9C illustrates example plots;

[0014] FIG. 10 illustrates an example of a system;

[0015] FIG. 11 illustrates an example of a system;

[0016] FIG. 12 illustrates an example of a system;

[0017] FIG. 13 illustrates an example of a method and an example of a system; and

[0018] FIG. 14 illustrates example components of a system and a networked system.DETAILED DESCRIPTION

[0019] The following description includes the best mode presently contemplated for practicing the described implementations. This description is not to be taken in a limiting sense, but rather is made merely for the purpose of describing the general principles of the implementations. The scope of the described implementations should be ascertained with reference to the issued claims.

[0020] Below, various types of environments, frameworks, workflows, data acquisition techniques, field equipment, field operations, etc., are described, which may involve use of a framework or frameworks, optionally during one or more field operations.

[0021] FIG. 1 shows an example of a system 100 that includes a workspace framework 110 that can provide for instantiation of, rendering of, interactions with, etc., a graphical user interface (GUI) 120. In the example of FIG. 1 , the GUI 120 can include graphical controls for computational frameworks (e.g., applications) 121 , projects 122, visualization 123, one or more other features 124, data access 125, and data storage 126.

[0022] In the example of FIG. 1 , the workspace framework 110 may be tailored to a particular geologic environment such as an example geologic environment 150. For example, the geologic environment 150 may include layers (e.g., stratification) that include a reservoir 151 and that may be intersected by a fault 153. A geologic environment 150 may be outfitted with a variety of sensors, detectors, actuators, etc. In such an environment, various types of equipment such as, for example, equipment 152 may include communication circuitry to receive and to transmit information, optionally with respect to one or more networks 155. Such information may include information associated with downhole equipment 154, which may be equipment to acquire information, to assist with resource recovery, etc.Other equipment 156 may be located remote from a wellsite and include sensing, detecting, emitting, or other circuitry. Such equipment may include storage and communication circuitry to store and to communicate data, instructions, etc. One or more satellites may be provided for purposes of communications, data acquisition, etc. For example, FIG. 1 shows a satellite 170 in communication with the network 155 that may be configured for communications, noting that the satellite may additionally or alternatively include circuitry for imagery (e.g., spatial, spectral, temporal, radiometric, etc.).

[0023] FIG. 1 also shows the geologic environment 150 as optionally including equipment 157 and 158 associated with a well that includes a substantially horizontal portion that may intersect with one or more fractures 159. For example, consider a well in a formation that may include natural fractures, artificial fractures (e.g., hydraulic fractures) or a combination of natural and artificial fractures. As an example, a well may be drilled for a reservoir that is laterally extensive. In such an example, lateral variations in properties, stresses, etc., may exist where an assessment of such variations may assist with planning, operations, etc., to develop a laterally extensive reservoir (e.g., via fracturing, injecting, extracting, etc.). As an example, the equipment 157 and / or 158 may include components, a system, systems, etc. for fracturing, seismic sensing, analysis of seismic data, assessment of one or more fractures, etc.

[0024] In the example of FIG. 1 , the GUI 120 shows some examples of computational frameworks, including the DRILLPLAN, PETREL, TECHLOG, PETROMOD, ECLIPSE, INTERSECT, KINETIX / VISAGE, and PIPESIM frameworks (SLB, Houston, Texas). One or more types of frameworks may be implemented within or in a manner operatively coupled to the DELFI environment, which is a secure, cognitive, cloud-based collaborative environment that integrates data and workflows with digital technologies, such as artificial intelligence (Al) and machine learning (ML). Such an environment can provide for operations that involve one or more frameworks. The DELFI environment may be referred to as the DELFI framework, which may be a framework of frameworks. The DELFI environment can include various other frameworks, which may operate using one or more types of models (e.g., simulation models, etc.).

[0025] The DRILLPLAN framework provides for digital well construction planning and includes features for automation of repetitive tasks and validation workflows, enabling improved quality drilling programs (e.g., digital drilling plans, etc.) to be produced quickly with assured coherency.

[0026] The DRILLOPS framework, which may be included in the system 100 of FIG. 1 , may execute a digital drilling plan and ensures plan adherence, while delivering goal-based automation. The DRILLOPS framework may generate activity plans automatically individual operations, whether they are monitored and / or controlled on the rig or in town. Automation may utilize data analysis and learning systems to assist and optimize tasks, such as, for example, setting ROP to drilling a stand. A preset menu of automatable drilling tasks may be rendered, and, using data analysis and models, a plan may be executed in a manner to achieve a specified goal, where, for example, measurements may be utilized for calibration. The DRILLOPS framework provides flexibility to modify and replan activities dynamically, for example, based on a live appraisal of various factors (e.g., equipment, personnel, and supplies). Well construction activities (e.g., tripping, drilling, cementing, etc.) may be continually monitored and dynamically updated using feedback from operational activities. The DRILLOPS framework may provide for various levels of automation based on planning and / or re-planning (e.g., via the DRILLPLAN framework), feedback, etc.

[0027] The PETREL framework can be part of the DELFI cognitive exploration and production (E&P) environment (SLB, Houston, Texas, referred to as the DELFI environment) for utilization in geosciences and geoengineering, for example, to analyze subsurface data from exploration to production of fluid from a reservoir.

[0028] The TECHLOG framework can handle and process field and laboratory data for a variety of geologic environments (e.g., deepwater exploration, shale, etc.). The TECHLOG framework can structure wellbore data for analyses, planning, etc. As an example, the TECHLOG framework may be coupled to one or more ML models for purposes of generation of output, training, etc.

[0029] The PETROMOD framework provides petroleum systems modeling capabilities that can combine one or more of seismic, well, and geological information to model the evolution of a sedimentary basin. The PETROMOD framework can predict if, and how, a reservoir has been charged with hydrocarbons,including the source and timing of hydrocarbon generation, migration routes, quantities, and hydrocarbon type in the subsurface or at surface conditions.

[0030] The ECLIPSE framework provides a reservoir simulator with numerical solvers for prediction of dynamic behavior for various types of reservoirs and development schemes.

[0031] The INTERSECT framework provides a high-resolution reservoir simulator for simulation of geological features and quantification of uncertainties, for example, by creating production scenarios and, with the integration of precise models of the surface facilities and field operations, the INTERSECT framework can produce results, which may be continuously updated by real-time data exchanges (e.g., from one or more types of data acquisition equipment in the field that can acquire data during one or more types of field operations, etc.). The INTERSECT framework can provide completion configurations for complex wells where such configurations can be built in the field, can provide detailed chemical-enhanced-oil- recovery (EOR) formulations where such formulations can be implemented in the field, can analyze application of steam injection and other thermal EOR techniques for implementation in the field, advanced production controls in terms of reservoir coupling and flexible field management, and flexibility to script customized solutions for improved modeling and field management control. The INTERSECT framework, as with the other example frameworks, may be utilized as part of the DELFI environment, for example, for rapid simulation of multiple concurrent cases.

[0032] The KINETIX framework provides for reservoir-centric stimulation-to- production analyses that can integrate geology, petrophysics, completion engineering, reservoir engineering, and geomechanics, for example, to provide for optimized completion and fracturing designs for a well, a pad, or a field. The KINETIX framework can be operatively coupled to and / or integrated with features of the PETREL framework (e.g., within the DELFI environment). As to the VISAGE framework it can be part of or otherwise operatively coupled to the KINETIX framework.

[0033] The VISAGE framework includes finite element numerical solvers that may provide simulation results such as, for example, results as to compaction and subsidence of a geologic environment, well and completion integrity in a geologic environment, cap-rock and fault-seal integrity in a geologic environment, fracturebehavior in a geologic environment, thermal recovery in a geologic environment, CO2 disposal, etc.

[0034] As an example, the KINETIX framework can provide for analyses from 1 D logs and simple geometric completions to 3D mechanical and petrophysical models coupled with the INTERSECT framework high-resolution reservoir simulator and VISAGE framework finite-element geomechanics simulator. The KINETIX framework can provide automated parallel processing using cloud platform resources and can provide for rapid assessment of well spacing, completion, and treatment design choices, enabling exploration of many scenarios in a relatively rapid manner (e.g., via provisioning of cloud platform resources).

[0035] The PIPESIM simulator includes solvers that may provide simulation results such as, for example, multiphase flow results (e.g., from a reservoir to a wellhead and beyond, etc.), flowline and surface facility performance, etc. The PIPESIM simulator may be integrated, for example, with the AVOCET production operations framework (SLB, Houston, Texas). The PIPESIM simulator may be an optimizer that can optimize one or more operational scenarios at least in part via simulation of physical phenomena.

[0036] As an example, a system may include or be operatively coupled to the SYMMETRY framework (SLB, Houston, Texas). The SYMMETRY framework provides features to model process workflows integrating facilities, process units with pipelines, fluid networks, flares, and safety systems, while ensuring consistent thermodynamics and fluid characterization across a system. Such an approach provides for optimizing processes in upstream, midstream and downstream sectors, which may be utilized for maximizing return, minimizing CAPEX, assessing emissions, assessing resource inputs, etc.

[0037] The aforementioned DELFI environment provides various features for workflows as to subsurface analysis, planning, construction and production, for example, as illustrated in the workspace framework 110. As shown in FIG. 1 , outputs from the workspace framework 110 can be utilized for directing, controlling, etc., one or more processes in the geologic environment 150, and feedback 160 can be received via one or more interfaces in one or more forms (e.g., acquired data as to operational conditions, equipment conditions, environment conditions, etc.).

[0038] In the example of FIG. 1 , the visualization features 123 may be implemented via the workspace framework 110, for example, to perform tasks as associated with one or more of subsurface regions, planning operations, constructing wells and / or surface fluid networks, and producing from a reservoir.

[0039] Visualization features may provide for visualization of various earth models, properties, etc., in one or more dimensions. As an example, visualization features may include one or more control features for control of equipment, which can include, for example, field equipment that can perform one or more field operations. A workflow may utilize one or more frameworks to generate information that can be utilized to control one or more types of field equipment (e.g., drilling equipment, wireline equipment, fracturing equipment, etc.).

[0040] FIG. 2 shows an example of a system 200 that can be operatively coupled to one or more databases, data streams, etc. For example, one or more pieces of field equipment, laboratory equipment, computing equipment (e.g., local and / or remote), etc., can provide and / or generate data that may be utilized in the system 200.

[0041] As shown, the system 200 can include a geological / geophysical data block 210, a surface models block 220 (e.g., for one or more structural models), a volume modules block 230, an applications block 240, a numerical processing block 250 and an operational decision block 260. As shown in the example of FIG. 2, the geological / geophysical data block 210 can include data from well tops or drill holes 212, data from seismic interpretation 214, data from outcrop interpretation and optionally data from geological knowledge. As an example, the geological / geophysical data block 210 can include data from digital images, which can include digital images of cores, cuttings, cavings, outcrops, etc. As to the surface models block 220, it may provide for creation, editing, etc. of one or more surface models based on, for example, one or more of fault surfaces 222, horizon surfaces 224 and optionally topological relationships 226. As to the volume models block 230, it may provide for creation, editing, etc. of one or more volume models based on, for example, one or more of boundary representations 232 (e.g., to form a watertight model), structured grids 234 and unstructured meshes 236.

[0042] As shown in the example of FIG. 2, the system 200 may allow for implementing one or more workflows, for example, where data of the data block 210are used to create, edit, etc. one or more surface models of the surface models block 220, which may be used to create, edit, etc. one or more volume models of the volume models block 230. As indicated in the example of FIG. 2, the surface models block 220 may provide one or more structural models, which may be input to the applications block 240. For example, such a structural model may be provided to one or more applications, optionally without performing one or more processes of the volume models block 230 (e.g., for purposes of numerical processing by the numerical processing block 250). Accordingly, the system 200 may be suitable for one or more workflows for structural modeling (e.g., optionally without performing numerical processing per the numerical processing block 250).

[0043] As to the applications block 240, it may include applications such as a well prognosis application 242, a reserve calculation application 244 and a well stability assessment application 246. As to the numerical processing block 250, it may include a process for seismic velocity modeling 251 followed by seismic processing 252, a process for facies and petrophysical property interpolation 253 followed by flow simulation 254, and a process for geomechanical simulation 255 followed by geochemical simulation 256. As indicated, as an example, a workflow may proceed from the volume models block 230 to the numerical processing block 250 and then to the applications block 240 and / or to the operational decision block 260. As another example, a workflow may proceed from the surface models block 220 to the applications block 240 and then to the operational decisions block 260 (e.g., consider an application that operates using a structural model).

[0044] In the example of FIG. 2, the operational decisions block 260 may include a seismic survey design process 261 , a well rate adjustment process 252, a well trajectory planning process 263, a well completion planning process 264 and a process for one or more prospects, for example, to decide whether to explore, develop, abandon, etc. a prospect.

[0045] Referring again to the data block 210, the well tops or drill hole data 212 may include spatial localization, and optionally surface dip, of an interface between two geological formations or of a subsurface discontinuity such as a geological fault; the seismic interpretation data 214 may include a set of points, lines or surface patches interpreted from seismic reflection data, and representing interfaces between media (e.g., geological formations in which seismic wave velocitydiffers) or subsurface discontinuities; the outcrop interpretation data 216 may include a set of lines or points, optionally associated with measured dip, representing boundaries between geological formations or geological faults, as interpreted on the earth surface; and the geological knowledge data 218 may include, for example knowledge of the paleo-tectonic and sedimentary evolution of a region.

[0046] As to a structural model, it may be, for example, a set of gridded or meshed surfaces representing one or more interfaces between geological formations (e.g., horizon surfaces) or mechanical discontinuities (fault surfaces) in the subsurface. As an example, a structural model may include some information about one or more topological relationships between surfaces (e.g., fault A truncates fault B, fault B intersects fault C, etc.).

[0047] As to the one or more boundary representations 232, they may include a numerical representation in which a subsurface model is partitioned into various closed units representing geological layers and fault blocks where an individual unit may be defined by its boundary and, optionally, by a set of internal boundaries such as fault surfaces.

[0048] As to the one or more structured grids 234, it may include a grid that partitions a volume of interest into different elementary volumes (cells), for example, that may be indexed according to a pre-defined, repeating pattern. As to the one or more unstructured meshes 236, it may include a mesh that partitions a volume of interest into different elementary volumes, for example, that may not be readily indexed following a pre-defined, repeating pattern (e.g., consider a Cartesian cube with indexes I, J, and K, along x, y, and z axes).

[0049] As to the seismic velocity modeling 251 , it may include calculation of velocity of propagation of seismic waves (e.g., where seismic velocity depends on type of seismic wave and on direction of propagation of the wave). As to the seismic processing 252, it may include a set of processes allowing identification of localization of seismic reflectors in space, physical characteristics of the rocks in between these reflectors, etc.

[0050] As to the facies and petrophysical property interpolation 253, it may include an assessment of type of rocks and of their petrophysical properties (e.g., porosity, permeability), for example, optionally in areas not sampled by well logs orcoring. As an example, such an interpolation may be constrained by interpretations from log and core data, and by prior geological knowledge.

[0051] As to the flow simulation 254, as an example, it may include simulation of flow of hydro-carbons in the subsurface, for example, through geological times (e.g., in the context of petroleum systems modeling, when trying to predict the presence and quality of oil in an un-drilled formation) or during the exploitation of a hydrocarbon reservoir (e.g., when some fluids are pumped from or into the reservoir).

[0052] As to geomechanical simulation 255, it may include simulation of the deformation of rocks under boundary conditions. Such a simulation may be used, for example, to assess compaction of a reservoir (e.g., associated with its depletion, when hydrocarbons are pumped from the porous and deformable rock that composes the reservoir). As an example, a geomechanical simulation may be used for a variety of purposes such as, for example, prediction of fracturing, reconstruction of the paleo-geometries of the reservoir as they were prior to tectonic deformations, etc.

[0053] As to geochemical simulation 256, such a simulation may simulate evolution of hydrocarbon formation and composition through geological history (e.g., to assess the likelihood of oil accumulation in a particular subterranean formation while exploring new prospects).

[0054] As to the various applications of the applications block 240, the well prognosis application 242 may include predicting type and characteristics of geological formations that may be encountered by a drill bit, and location where such rocks may be encountered (e.g., before a well is drilled); the reserve calculations application 244 may include assessing total amount of hydrocarbons or ore material present in a subsurface environment (e.g., and estimates of which proportion can be recovered, given a set of economic and technical constraints); and the well stability assessment application 246 may include estimating risk that a well, already drilled or to-be-drilled, will collapse or be damaged due underground stress.

[0055] As to the operational decision block 260, the seismic survey design process 261 may include deciding where to place seismic sources and receivers to optimize the coverage and quality of the collected seismic information while minimizing cost of acquisition; the well rate adjustment process 262 may includecontrolling injection and production well schedules and rates (e.g., to maximize recovery and production); the well trajectory planning process 263 may include designing a well trajectory to maximize potential recovery and production while minimizing drilling risks and costs; the well trajectory planning process 264 may include selecting proper well tubing, casing and completion (e.g., to meet expected production or injection targets in specified reservoir formations); and the prospect process 265 may include decision making, in an exploration context, to continue exploring, start producing or abandon prospects (e.g., based on an integrated assessment of technical and financial risks against expected benefits).

[0056] The system 200 can include and / or can be operatively coupled to a system such as the system 100 of FIG. 1 . For example, the workspace framework 110 may provide for instantiation of, rendering of, interactions with, etc., the graphical user interface (GUI) 120 to perform one or more actions as to the system 200. In such an example, access may be provided to one or more frameworks (e.g., DRILLPLAN, DRILLOPS, PETREL, TECHLOG, PETROMOD, ECLIPSE, INTERSECT, KINETIX / VISAGE, PIPESIM, SYMMETRY, etc.). One or more frameworks may provide for geo data acquisition as in block 210, for structural modeling as in block 220, for volume modeling as in block 230, for running an application as in block 240, for numerical processing as in block 250, for operational decision making as in block 260, etc.

[0057] As an example, the system 200 may provide for monitoring data, which can include geo data per the geo data block 210. In various examples, geo data may be acquired during one or more operations. For example, consider acquiring geo data during drilling operations via downhole equipment and / or surface equipment. As an example, the operational decision block 260 can include capabilities for monitoring, analyzing, etc., such data for purposes of making one or more operational decisions, which may include controlling equipment, revising operations, revising a plan, etc. In such an example, data may be fed into the system 200 at one or more points where the quality of the data may be of particular interest. For example, data quality may be characterized by one or more metrics where data quality may provide indications as to trust, probabilities, etc., which may be germane to operational decision making and / or other decision making.

[0058] As an example, the system 200 may include one or more features for operations involving storage of fluid in a subsurface region and / or production of fluid from a subsurface region. For example, consider operations that may involve fluid that may include one or more of hydrogen, carbon, oxygen, sulfur, etc. As an example, the system 200 may include one or more ML models, for example, consider one or more ML models for use in one or more types of workflows. As an example, the system 200 may include one or more hybrid models, which may include one or more physics-based portion and one or more data-based portions. As an example, the system 200 may include one or more physics-informed ML models and / or implementation of one or more physics-informed ML techniques.

[0059] As explained, fluid may be stored in a subsurface region where such fluid may include hydrogen. For example, consider an operation that involves injecting hydrogen into one or more subsurface reservoirs where, for example, the hydrogen may be produced from one or more of the one or more subsurface reservoirs during one or more times of higher demand for hydrogen.

[0060] Upon injection, hydrogen may contact one or more host fluids such as, for example, brine and in-place gases such as one or more of methane, carbon dioxide, and hydrogen sulfide. As explained with respect to the system 200, various subsurface and / or surface processes may be simulated, optimized, controlled, etc.

[0061] As an example, a system may include a framework or be operatively coupled to a framework that can provide for simulation and / or optimization of one or more processes associated with storage of fluid and / or production of a stored fluid. For example, consider a framework that can provide for accurate modeling and prediction of hydrogen solubility in an aqueous phase as well as, for example, water content of a gas phase. As an example, such a framework may provide for control of one or more processes, pieces of field equipment, etc.

[0062] As an example, a framework may provide for implementation of a physics-informed ML approach to generate values for solubility and water content of hydrogen / pure water and hydrogen / brine systems. In various trials, such a framework, as evaluated using experimental solubility and water content data, achieved acceptable accuracy over temperature and pressure ranges of interest such as, for example, temperature up to approximately 472.15 K and pressure up to approximately 500 bar, along with salt concentration up to approximately 5 mol / kg.As an example, an approximate value may be within plus or minus 10 percent, 5 percent, or 2.5 percent, or less (e.g., 1 percent), which may depend on circumstances. Such a framework can provide for generating output for one or more purposes, which can include simulation of physical phenomena, planning of operations, selection of subsurface regions, control of operations, optimization of operations, etc.

[0063] As explained, hydrogen may be stored in a subsurface region and / or produced from a subsurface region. As an example, hydrogen may serve as a fuel that may replace and / or supplement one or more fossil fuels, which in various instances may help to mitigate environmental concerns associated with fossil fuels. Various fossil fuels may exist in reservoirs as part of natural geologic processes. For example, consider crude oil, methane, etc., where production of such fluids may be a concern. As to hydrogen, while there may be stores of natural hydrogen in various subsurface regions that may be amenable to production of such natural hydrogen, various workflows may involve injection of hydrogen for storage thereof, where such stored hydrogen may later be produced and utilized as a fuel. As to natural reserves of subsurface hydrogen, they may stem from water-rock reactions deep within the Earth that may generate hydrogen where such generated hydrogen may then percolate up through the crust and sometimes accumulates in underground traps. As to production of natural hydrogen from such traps, one or more production techniques may involve injection of fluid to promote production of natural hydrogen. In such an approach, interactions between fluids may be of interest where, for example, a framework may provide for implementation of a physics-informed ML approach to generate values for solubility and water content of hydrogen / pure water and hydrogen / brine systems. Hence, such a framework may be suitable for implementation as to production of stored hydrogen (e.g., as previously injected) and / or as to production of natural hydrogen (e.g., as naturally accumulated). In either of such scenarios, hydrogen interactions with aqueous fluid may occur where generation of values for solubility and water content of hydrogen / pure water and hydrogen / brine systems may be of interest (e.g., for planning, developing, controlling, etc.).

[0064] As an example, a scenario for storage of hydrogen may involve one or more technologies and / or techniques employed for storage of liquified natural gas(LNG). For example, injection equipment, injection wells, etc., may be employed for storage of hydrogen and / or LNG. As to hydrogen, as mentioned, hydrogen solubility in water and brine can be factors as well as, for example, water content of a hydrogen-rich phase. Such factors may be parameters that can be utilized in simulation and assessment of hydrogen storage in one or more geological formations such as, for example, one or more depleted oil and gas reservoirs and / or one or more saline aquifers.

[0065] As to solubility, Henry’s law may be utilized, which is a gas law that states that the amount of dissolved gas in a liquid is directly proportional to its partial pressure above the liquid. In such an approach, a proportionality factor may be referred to as a Henry’s law constant. An example of where Henry’s law is at play is the depth-dependent dissolution of oxygen and nitrogen in the blood of underwater divers that changes during decompression, leading to decompression sickness. As another example, consider bottled carbonated beverages that contain dissolved carbon dioxide. Before opening a bottle, gas above the beverage in the bottle may be nearly pure carbon dioxide, at a pressure higher than atmospheric pressure.After the bottle is opened, this gas escapes, moving the partial pressure of carbon dioxide above the liquid to be much lower, resulting in degassing as the dissolved carbon dioxide comes out of the beverage.

[0066] As to hydrogen solubility, various models tend to use Equations of State (EoS or EOS) or thermodynamic correlations to compute hydrogen solubility in water and brine. Implementation of a cubic EOS for water-containing systems often render inaccurate results. This is the point where cubic-plus-association EOS have been introduced to improve the accuracy of these models. However, these models tend to be computationally expensive, especially for large-scale reservoir simulations. For example, Rahbari et al. modeled the H2-H2O system using a molecular simulation technique that showed better performance over the Peng- Robinson equation of state (PR-EOS).

[0067] The PR-EOS was developed in an effort to satisfy the following goals: the parameters should be expressible in terms of the critical properties and the acentric factor; the model should provide reasonable accuracy near the critical point, particularly for calculations of the compressibility factor and liquid density; the mixing rules should not employ more than a single binary interaction parameter, whichshould be independent of temperature, pressure, and composition; and the equation should be applicable to all calculations of all fluid properties in natural gas processes.

[0068] Rahbari et al. noted that hydrogen, because of its relatively low volumetric energy density, should be compressed for practical storage and transportation purposes and that electrochemical hydrogen compressors (EHCs) may be utilized that (e.g., capable of compressing hydrogen up to 1000 bar). In an EHC, compressed hydrogen is saturated with water, where the maximum water content in gaseous hydrogen may meet a fuel requirement such as, for example, the International Organization for Standardization (ISO) when refueling fuel cell electric vehicles. Specifically, the ISO 14687-2:2012 standard limits the water concentration in hydrogen gas to 5 pmol water per mol hydrogen fuel mixture. Thus, knowledge of the vapor liquid equilibrium of H2O-H2 mixtures can be helpful in designing a method to remove H2O from compressed H2. As mentioned, Rahbari et al. turned to molecular simulation and thermodynamic modeling to study the phase coexistence of the H2O-H2 system and to show that the PR-EOS and the Soave Redlich-Kwong EOS with van der Waals mixing rules fail to accurately predict the equilibrium coexistence compositions of the liquid and gas phase, with or without fitted binary interaction parameters. Rahbari et al. demonstrated that solubility of water in compressed hydrogen may be adequately predicted using force-field-based molecular simulations. However, as mentioned, such simulations can be resource intensive as to time and / or computational power. Hence, such simulations may introduce latencies, demand for more resources (e.g., provisioning of compute, memory, etc.), etc., which may make such simulations impractical for integration into one or more types of workflows (e.g., workflows that may involve relatively lightweight computing devices, may involve time constraints, may be limited in terms of transmission bandwidth, etc.). As an example, a framework that implements an ML approach, which may be a physics-informed ML approach, may provide for suitable accuracy and efficiency, thereby enabling one or more workflows that are impractical using the simulation-based approach of Rahbari et al.

[0069] While some have turned to ML to compute phase equilibrium in chemical and petroleum engineering fields and hydrogen solubility in water (e.g., using artificial neural networks (ANN)), these approaches use input parameters suchas T, P, and salt (NaCI) concentration to calculate the hydrogen solubility in the liquid phase. Hence, these approaches suffer due to their limitations to pure hydrogen in the gas phase. Accordingly, such models cannot capture the effect of impurities such as, for example, one or more of carbon dioxide (CO2), hydrogen sulfide (H2S), and methane (CH4) in the gas phase.

[0070] As an example, a physics-informed machine learning (PIML) approach may be utilized as an efficient technique to integrate machine learning and physics behind relevant phenomena. As an example, a framework can implement a physics- informed machine learning (PIML) technique to generate values for hydrogen solubility in water and brine as well as hydrogen water content. In various trials, a PIML, neural network-based approach provides an accurate thermodynamic technique that is less computationally expensive compared to current EOS techniques that implement iterative temperature and pressure (TP) flash calculations. As an example, a PIML approach may model gas phase properties using a suitable EOS (e.g., PR-EOS) to benefit from its accurate gas properties. As to complexities of the aqueous phase, these may be modeled using a neural network model and / or one or more other suitable ML models (e.g., data-driven models, etc.). As an example, a framework may provide for flexibility and extensibility. For example, a PIML approach may extend the capability of a model for impure hydrogen systems (e.g., CH4, CO2, H2S, etc.), which may thereby improve capabilities over a pure neural network approach.

[0071] As an example, use of the PR-EOS to model the gas phase allows for leveraging the accuracy of this EOS for calculating the gas phase properties such as, for example, density and fugacity. In chemical thermodynamics, the fugacity of a real gas is an effective partial pressure that can replace a mechanical partial pressure in a computation of chemical equilibrium. The fugacity can be equal to the pressure of an ideal gas that has the same temperature and molar Gibbs free energy as the real gas. Fugacities tend to be determined experimentally or estimated from various models such as a Van der Waals gas that are closer to reality than an ideal gas. The real gas pressure and fugacity are related through the dimensionless fugacity coefficient <p.

[0072] As mentioned, the effect of the impurities in the hydrogen phase such as, for example, CO2 and CH4, may be captured by the PR-EOS. One or more ofsuch components may be present in one or more geological formations and may be contacted (e.g., mixed) by injected hydrogen. The effect of salinity on gas-phase water content and hydrogen solubility in the aqueous phase may also be captured using an ML model (e.g., consider a neural network model, etc.). A proposed model may be used to calculate water content of the hydrogen phase as well as hydrogen solubilities in pure water and saline over a relatively wide range of T and P, which may be of interest in scenarios that may involve geological hydrogen storage (e.g., consider T up to approx. 473.15 K and P up to approx. 500 bar). As an example, a model may provide for implementation of a noniterative approach to computing mutual solubilities. Such a capability can make the model favorable for large-scale models like hydrogen storage in subsurface geological reservoirs.

[0073] Where an ML model is or includes a neural network, the ML model may be implemented as part of a physics-informed neural network (PINN) that may be utilized for generating values for water content of the hydrogen gas phase, as well as hydrogen solubility in pure and saline waters for hydrogen storage in geological formations. Such an approach can allow for modeling to provide reliable results for different operational conditions, such as, for example, conditions involving the presence of one or more other components (e.g., CO2, CH4, etc.) in the gas phase. Such a model may enhance viability of and / or efficiency of one or more workflows, which may include workflows involving engineering calculations and simulations of hydrogen storage in geological containment systems.

[0074] As an example, a framework may involve modeling with fugacity equality for species in equilibrium. For example, consider hydrogen and water as follows: ftl= f9(D where / ) shows the fugacity of species i and the superscripts I and g show the liquid and gas phases, respectively.

[0075] As an example, fugacity of the gas phase can be written as: ft9= yt<PtP (2)where shows the mole fraction, p is the fugacity coefficient and P denotes the system pressure. As explained, the gas phase may be modeled using an approach EOS (e.g., consider the PR-EOS). Accuracy and simplicity of the PR-EOS has been demonstrated for gas-phase computations, particularly in various P and T ranges of interest in various geological storage scenarios for hydrogen.

[0076] As an example, the liquid phase may be modeled using Henry’s law, as given by: ft3= XiYiHi (3) where x, y and H represent mole fraction, activity coefficient, and Henry’s constant of species i in the liquid phase, respectively.

[0077] As an example, an ANN may be utilized as an ML model to model the liquid phase. The aqueous phase tends to be relatively complex and challenging to model using conventional EOSs, and can demand experimental solubility and density data for tuning. Therefore, a framework may implement one or more ANNs to correlate Henry’s constant and activity coefficient of the liquid phase. In such an example, an ANN with two inputs (T and P) and one output (H) may be used for generation of a value for Henry’s constant. To generate a value of the activity coefficient, as an example, an ANN may be utilized that includes three inputs: T, anion concentration, and cation concentration.

[0078] As an example, a procedure for computing hydrogen solubility in the aqueous phase and water content of the gas phase may include various actions. For example, consider a workflow that first computes an initial guess for the water mole fraction in the gas phase as the ratio of saturation pressure of water to total pressure Psat / P), followed by computing the fugacity coefficients for the hydrogen and water in the gas phase. In such a procedure, the water mole fraction in the gas phase may be computed as follows:

[0079] As an example, a procedure may ignore the hydrogen mole fraction in the liquid phase when using Eqn. (4) because the range of hydrogen mole fraction in the aqueous phase may tend to be relatively small (e.g., less than approximately 0.01 %). In such an approach, ignoring the hydrogen mole fraction does not necessarily introduce unacceptable error in computing the water content of the hydrogen gas phase. As an example, the mole fraction of water in the liquid phase can be assumed as 1 - xsaUfor hydrogen-water systems. After computing the gas phase composition, a process may then compute the hydrogen mole fraction in the liquid phase, for example, as follows:(1 yn2o)P(PH2(5) xH= -2 HH2YH2

[0080] As explained, Henry’s constant and the activity coefficient may be computed from tuned ANN models. To tune one or more ANN models, experimental data for water content of hydrogen as well as the hydrogen solubility in pure water and salty water may be utilized. For example, a computational framework, which may utilize one or more libraries (e.g., via application programming interfaces (APIs), packages, etc.), may provide for acquisition of experimental data, as may be acquired using one or more types of laboratory equipment, one or more databases, etc., and provide for tuning one or more ML (e.g., ANN, etc.) models, which may be via one or more techniques. As an example, tuning may involve weighting of weights, adjustment of one or more hyperparameters, etc. As an example, a computational framework may provide for rendering of one or more graphical user interfaces for human-machine interactions that may provide for acquisition of data, tuning, etc.

[0081] FIG. 3 shows an example of an architecture 300 for various components of a framework. For example, the architecture 300 can include components for input 310, an EOS 320, an ML model 332 for equilibrium constant (K such as for water), an ML model 334 for Henry’s constant (H such as for hydrogen(e.g., HH2)), and an ML model 336 for activity coefficient (y such as for hydrogen), a fugacity computation 340 (<p such as for water and for hydrogen), and output 350. In such an example, the input component 310 may provide input as to mole fraction of hydrogen (yH2), pressure (P), temperature (T), and molarity of bulk brine (e.g., salt in aqueous phase, such as mSait) and the output component 350 may provide for output of values of water mole fraction in a liquid phase (e.g., water as a component in the liquid phase, where Xi is XH20) and water mole fraction in a gas phase (e.g., water as a component in the gas phase, where yi is yH2o). As to the subscript i, it can represent a component where multiple components can be present (see, e.g., summation equation for the fractions of each phase in the fugacity computation component 340). As an example, the output component 350 can provide for output of mole fractions where each component has a mole fraction output for the liquid phase and for the gas phase, noting that in some instances, a component may exist in either the liquid phase or the gas phase but not both the liquid phase and the gas phase.

[0082] While the example of FIG. 3 includes water, hydrogen and salt, the architecture 300 may be operable for a system that includes water and hydrogen, for example, without salt. For example, if salt concentration is sufficiently low, it may be neglected. As an example, a framework may provide for automatically comparing a salt concentration to a threshold and determining how to operate the framework, for example, with or without salt, as may be via selection of one or more ML models, one or more physics-informed models, etc.). In the example of FIG. 3, the architecture 300 may handle scenarios that involve a liquid phase and a potential existence of gas phase (e.g., to determine if gas may form).

[0083] In the example of FIG. 3, the EOS component 320 may provide for a physics-informed approach to generation of output per the output component 350 where, for example, the EOS component 320 may generate fugacity coefficients that are utilized by the fugacity computation component 340 along with outputs of the ML models 332, 334, and 336.

[0084] As an example, the architecture 300 may be for a PINN that can compute mutual solubilities of hydrogen-water and hydrogen-brine systems. As an example, the architecture 300 may be flexible and / or extensible. For example, the type of EOS may be selectable and / or the type of ML model or ML models may beselectable; noting that the number of ML models may be selectable. For example, where one or more additional fluids are present, one or more ML models may be introduced.

[0085] As to the EOS component 320, consider, as an example, using a Peng-Robinson (PR) EOS: n RT aP~ V — b ~ V(V + b) + b(V - b)

[0086] As to the fugacity computation component 340, consider, for example, using one or more of the following equations:

[0087] As explained, underground hydrogen storage may be implemented as a storage method in which hydrogen is injected into one or more subsurface reservoirs, for example, for later produced (e.g., during times of higher demand, etc.). Injected hydrogen may contact host fluid such as, for example, brine and in- place gases such as, for example, one or more of methane, carbon dioxide, and hydrogen sulfide. Planning, development, operation, simulation, optimization, etc., of relevant processes demands a relatively accurate and expeditious model for predicting hydrogen solubility in an aqueous phase as well as water content of a gas phase. As explained, a physics-informed machine-learning (PIML) approach may be applied, which may be implemented as a computational framework that may be instructed to computer solubility and water content of one or more systems (e.g., hydrogen / pure water, hydrogen / brine, etc.). As explained, one or more features of the architecture 300 of FIG. 3 may be employed as part of a PIML approach where, for example, a gas phase may be modeled using classic thermodynamics (e.g., an equation of state (EoS)) and a liquid phase may be modeled using one or more artificial neural networks (ANNs). As an example, a PIML approach may be implemented in combination with one or more types of frameworks. For example,consider the PETREL framework for various processes, the SYMMETRY framework for various processes, etc., where in one or more of such processes, hydrogen contacts water, which may be brine.

[0088] As an example, a framework may provide for simulation of hydrogen and / or CO2 storage in one or more underground sites where the framework provides relatively accurate and computationally efficient thermodynamic modeling features to predict, for example, gas solubility in an aqueous phase. As explained, a PIML approach may compute gas (e.g., hydrogen, etc.) solubility in water in a manner that improves the capability of simulation of hydrogen and / or CO2 storage scenarios. As explained, such an approach may be implemented in combination with a reservoir simulator and / or process simulator, for example, to simulate hydrogen and / or CO2 storage in underground reservoirs and / or processes that may be associated therewith.

[0089] FIG. 4 shows an example formulation for a PR-EOS 400, which can include various parameters. As an example, the PR-EOS 400 may be utilized by theEOS component 320 of the architecture 300 of FIG. 3. For example, consider the following example equation:wherekPR= 0.37464 + 1.524226a) - 0.26992a and where Tris the reduced temperature T / Tc, the parameters a0and b are computed from the critical properties and kPRis computed from the accentric factor O).

[0090] As an example, a framework may provide for normalizing ANN input and output parameters, for example, within a range from 0.0 to 1 .0. As an example, an equation such as, for example, Eqn. (6), below, may be implemented to normalizeinput parameters such as, for example, T, P, and NaCI concentration and output parameter for each tuned ANN.

[0091] In an example implementation, TENSORFLOW and KERAS were utilized (e.g., TENSORFLOW2.12.0) to create, fit, and run ANN models. In this example, dense layers with a leaky rectified linear unit, or leaky ReLU, activation model were used for the ANN models (see, e.g., the ANNs 332, 334, and 336 of FIG. 3). In the example implementation, for H2O, HH2, and yH2 the layers nodes were (8, 10, 10, 10, 4, 1 ), (2, 8, 20, 40, 20, 10, 1 ) and (2, 8, 20, 10, 1 ), respectively. A root mean square propagation (RMSprop) optimizer was used to optimize the layers. PYTHON version 3.10 was utilized for purposes of programming and executing models. As indicated, the ML models in the architecture 300 may differ. As an example, by tailoring models to specific tasks, a model may be made more robust and an architecture may be made modular where, for example, one or more modules may be added, removed, turned on, turned off, modified, etc. As an example, a model may be generated (e.g., trained) using available data that may be specific to a particular type of input and output.

[0092] FIG. 5A and FIG. 5B shows example plots that include comparative data. In FIG. 5A and FIG. 5B, the aforementioned example PINN model was implemented to regenerate experimental data of water content of the hydrogen phase. The plots show comparisons between experimental and calculated water content data. Specifically, the plots show comparisons between experimental and calculated water content of the hydrogen phase where experimental data were collected from Gillespie and Wilson, Ugrozov, Maslennikova et al., Bartlett, and Devaney et al. In FIG. 5A and FIG. 5B, each of the plots shows water mole fraction versus pressure for comparisons of computed data and experimental data.

[0093] As an example, accurate water content prediction is beneficial to have more accurate estimates of the stored amount of hydrogen as well as the amount of the produced hydrogen and water during production cycles. In various scenarios, water content of a hydrogen-rich phase increases by increasing T and decreasing P.In an example implementation of a PINN, experimental water content data of hydrogen at temperatures up to approx. 473.15 K were collected and used for model tuning / evaluation. Referring to the plots of FIG. 5A and FIG. 5B, acceptable agreement is achieved between experimental and computed water content data. More specifically, water content data reported by Gillespie and Wilson were reproduced with an absolute average relative deviation (AARD) of less than 6.9%. This experimental dataset is from a research report by the Gas Processing Association and is considered some of the most reliable and cited water content data of the hydrogen-water system in the open literature.

[0094] In an example implementation, a PINN model was used to compute the hydrogen solubility in pure and saline water where experimental solubility data in a wide range of T and P were collected and used for model evaluation.

[0095] FIG. 6 shows example plots 600 of solubility versus pressure for various temperatures, which are shown in detail in FIG. 7A, FIG. 7B, FIG. 7C, FIG. 7D, FIG. 7E, FIG. 7F, FIG. 7G, FIG. 7H, FIG. 7I, FIG. 7J, FIG. 7K, and FIG. 7L. The plots 600 show comparisons between experimental and computed hydrogen solubility in pure water. As can be seen in the plots 600, output values of the PINN framework are acceptable when compared to experimental solubility data. As shown in FIG. 6, via FIG. 7A to FIG. 7L, the experimental data of Wiebe and Gaddy have been reproduced by the PINN framework with an AARD of 5.3%, noting that this dataset covers temperatures up to 373.15 K and P up to 1000 bar, which tend to fall in an interest range of various hydrogen storage scenarios.

[0096] In an example implementation, experimental hydrogen solubility data of Gillespie and Wilson were also used to evaluate output of a PINN framework. Some deviation between experimental and computed data is observed for high-T data at very low pressures where the solubility of hydrogen is very low. In this region, there is a high uncertainty in the experimental measurements. An AARD of 8.5% was observed after excluding three data points (T=366.48 K, P = 3.4, 13.8, and 31 bar).

[0097] As noted, hydrogen solubility in pure water increases by increasing P. However, at a constant P, it decreases by increasing T until reaching the minimum solubility, then it increases by increasing T. The PINN framework has the capability of predicting this behavior. In particular, the effect of T on hydrogen solubility is shown to be more drastic at high-P conditions (e.g., P > approx. 150 bar). Such anability to predict behavior may be implemented within a control scheme. For example, consider a field control scheme that utilizes a controller (e.g., a control system, etc.) that may generate predictions for purposes of controlling field equipment. In such an example, control may consider predictions as to one or more parameters (e.g., solubility, temperature, pressure, etc.), which may be aligned with physics such as physical phenomena where hydrogen solubility in pure water increases by increasing P; decreases, at a constant P, by increasing T until reaching the minimum solubility; and then increases by increasing T.

[0098] FIG. 8 shows an example plot 800 of isobaric hydrogen solubility in pure water. In particular, solubility is shown versus temperature (e.g., 280 K to 440 K) for pressures of 50 bar, 200 bar, 300 bar, 500 bar and 800 bar.

[0099] As to implementation of a framework for practical applications of geological hydrogen storage, injected hydrogen may be in contact with the host fluid (e.g., brine). In such an example, the presence of NaCI in water tends to decrease the solubility of gases such as CO2, CP , and hydrogen. To compute the hydrogen solubility in saline water, the effect of dissolved NaCI may be captured using the hydrogen activity coefficient (y) (see, e.g., the ML model 336 of FIG. 3). As explained, values for this parameter may be computed using one or more ML models. For example, as shown in FIG. 3, input for an ML model may be T and NaCI molality where, for example, the effect of P may be neglected in an activity model. In an example implementation, data of Chabab et al. was used to evaluate model accuracy for prediction of hydrogen solubility in saline water.

[0100] FIG. 9A, FIG. 9B and FIG. 9C show example plots of solubility versus pressure for computed and experimental data values at various temperatures. The plots of FIG. 9A, FIG. 9B, and FIG. 9C show acceptable agreement between experimental and computed solubility data and an AARD of 5.6%. These data cover a range of NaCI concentrations from 1 to 5 mol / kg. As shown in FIG. 9A, FIG. 9B, and FIG. 9C, increasing the NaCI concentration in water results in lower hydrogen solubility in the aqueous phase. The plots show comparisons between experimental and computed hydrogen solubility in saline water where the experimental data are from Chabab et al.

[0101] Hydrogen storage in underground saline formations, caverns, and depleted oil and gas reservoirs plays an undebatable role in renewable energyutilization and energy decarbonization. For simulation of storage and production, process optimization, and reservoir estimations, an accurate and reliable model is desirable to compute hydrogen solubility in the aqueous phase and water content in the gas phase. As explained, a framework may implement a PIML approach that can capture different operating conditions such as T, P, NaCI salinity, and gas-phase composition. As to salts, one or more types of salts may be handled by a framework, for example, consider calcium salts, magnesium salts, potassium salts, etc. As an example, the architecture 300 of FIG. 3 may include one or more ML models to handle one or more types of salts.

[0102] As an example, a PIML approach may implement one or more ANNs, for example, to take a PINN approach. In various example implementations, a PINN approach to a PIML demonstrated acceptable computation of hydrogen solubility in pure and saline water and water content of the hydrogen-rich phase. As explained, a framework may model a gas phase using an acceptable EOS such as, for example, a PR-EOS where the framework may model a liquid phase using one or more ML models (e.g., one or more ANNs, etc.).

[0103] As shown in various example plots, in an example implementation, a PIML approach was evaluated using experimental data of gas-phase water content, and hydrogen solubility in pure water and brine. Results of the evaluation demonstrate accuracy of the PIML approach in regenerating experimental data over T and P ranges. As explained, a framework may provide for implementation of a PIML architecture in a manner that can provide for selecting a subsurface region, simulating phenomena in a subsurface region, planning operations, controlling operations, etc. As explained, operations may pertain to underground hydrogen storage where, for example, suitable temperature and pressure ranges exist, along with suitable concentrations of salt (e.g., NaCI). For example, consider T = 273.15 K to 473.15 K, P up to 500 bar, NaCI concentration up to 5 mol / kg. As an example, a framework may be implemented as to various types of scenarios, which can include scenarios for underground hydrogen storage and production.

[0104] FIG. 10 shows an example of a system 1000 that includes a facility 1010 that may provide H2 for storage and that may consume H2, equipment 1020 with conduits extending into a subsurface region 1030 and various processes 1040 where, for example, H2 can be sequestered for later production. In the example ofFIG. 10, the system 1000 can include and / or be operatively coupled to a gas / liquid framework 1015 that can provide for generating results (e.g., simulation results, controller results, etc.) for such field operations where the results can include results for interactions between injected gas and subsurface liquid. As shown, subsurface liquid may include salt. As explained, subsurface liquid may include one or more other types of constituents (e.g., CO2, carbonates, H2S, etc.). As an example, the framework 1015 may be in a control loop, which may be a feedback control loop. For example, sensor-based data may be fed to the framework 1015 to automatically drive the framework 1015 to control one or more pieces of field equipment. As explained, sensor-based data may include data as to temperature, pressure, solubility, salt concentration, etc., one or more of which may be germane to optimal control of a process or processes.

[0105] As an example, as to hydrogen injection, hydrogen gas may be injected into a subsurface region where it contacts at least water and may contact water and one or more salts. As an example, a subsurface region may be controlled as to temperature and / or pressure to adjust one or more factors that may provide for enhanced storage of hydrogen and / or enhanced production of hydrogen. As an example, a scenario may include producing hydrogen as gas via reduction in pressure upstream. In such an example, at least a portion of the hydrogen may be solubilized in water where upon a reduction in pressure occurs, the hydrogen can be released from the water such that the hydrogen can be produced. As an example, some amount of water may be produced upon production of hydrogen from a subsurface region where, for example, one or more separators may aim to separate out water from the mixture.

[0106] As an example, the framework 1015 may provide for generating output for one or more other types of injected, stored, and / or produced materials. For example, consider a scenario where clathrate hydrates may be utilized to store hydrogen and / or sequester CO2. Trapping of CO2 molecules in clathrate hydrates may be a controllable process that can provide a way to reduce CO2 levels in gas. As an example, clathrate hydrates may be utilized in separation of gases such as CO2 from gas (e.g., flue gases, desalination, etc.). For example, clathrate hydrates may be used in flue gases to separate CO2 by encouraging the formation of CO2 clathrate hydrate in a flue gas mixture. As to hydrate-based desalination, a processcan commence when clathrate hydrate forming agent is injected into seawater that has a surrounding temperature lower than clathrate hydrate forming temperature where such a condition promotes solidification and condensation of water molecules around the hydrate formers such that a slurry of clathrate ice and brine form. Once formed, brine may be separated from the slurry of clathrate ice and the clathrate melted via heat exchange with warmer surface water of the ocean.

[0107] As an example, a subsurface region may include trapped methane (CH4) where, for example, injected CO2 may cause release of trapped methane (CH4). As an example, methane may be trapped in clathrate hydrates and / or one or more other types of structures.

[0108] As an example, a framework may be integrated into a simulator such as a reservoir simulator (e.g., consider the INTERSECT simulator) and / or may be operatively coupled to a simulator such as a reservoir simulator, for example, via one or more application programing interfaces (APIs). As an example, a simulator such as a surface network simulator may be integrated with and / or may be operatively coupled to a framework. For example, consider the PIPESIM simulator. As an example, a reservoir simulator and a surface network simulator may be operatively coupled.

[0109] As an example, operation of a simulator may be improved upon integration with and / or coupling to a framework that provides for generation of mole fractions in gas and / or liquid phases. In such an example, the framework may provide for a PIML approach that is computationally less expensive than utilization of a pure EOS approach and / or a computationally expensive simulation approach. As explained, a framework may provide for handling a portion of modeling via a physicsbased approach and another portion of modeling via a data-driven ML-based approach. In such an example, the physics-based approach may utilize a relatively lightweight EOS such as, for example, the PR-EOS. As an example, an architecture such as the architecture 300 of FIG. 3 may provide for real-time or near real-time computation of values.

[0110] As to the PR-EOS being relatively lightweight, consider, in contrast, the GERG EOS (e.g., GERG-2008 EOS), which is a non-cubic EOS for natural gases and other mixtures of 21 natural gas components. Another example EOS is Soave’s modification of Benedict-Webb-Rubin (Soave-BWR) EOS, which despite its empiricalnature, provides accurate density description even around the critical point. The Soave-BWR EOS is much simpler than GERG-2008 and easier to handle and generalize to reservoir oil fluids. As an example, an EOS may be cubic (e.g., SRK, PR, etc.) or non-cubic (e.g., Soave-BWR, PC-SAFT, etc.).

[0111] FIG. 11 shows an example of a system 1100 that may be extensible, flexible, etc. As shown, the system 1100 can include a framework 1110 that may be interoperable with one or more other components, frameworks, etc. For example, consider a reservoir simulator 1120, a planner 1130, a surface network simulator 1140, a field controller 1150, a consumption demand monitor 1160, a production dynamics monitor 1170 and one or more other components 1180. As an example, production dynamics of hydrogen production and consumption dynamics of hydrogen may be linked where a framework may provide for controlling field operations to store hydrogen and / or produce stored hydrogen.

[0112] As an example, the system 1100 may include one or more features for controlling one or more simulators. For example, consider one or more application programming interfaces (APIs) that may trigger one or more actions responsive to receipt of an API call. For example, consider a framework that may provide for execution of code for one or more models, which may include one or more types of models, etc. As explained, one or more physics-informed models may be utilized, which may include one or more neural networks (see, e.g., FIG. 3, etc.). As an example, a simulator, as a computational system, may be controlled through interactions with a framework. For example, a simulator may be controlled as to its operation whereby the simulator may continue execution upon receipt of information from a framework, which may be information generated through use of one or more models (e.g., one or more ML models, etc.). As explained, an approach that utilizes one or more machine learning models may improve simulator operation, which may thereby enable improved workflows (e.g., planning, history matching, control, etc.).

[0113] As to APIs, as an example, a REST API approach may be utilized. For example, consider utilization of one or more REST APIs that communicate through HTTP requests to perform one or more functions like creating, reading, updating and deleting records (also known as CRLID) within a resource. For example, a REST API would use a GET request to retrieve a record; whereas, a POST request may create a new record. As an example, a framework may operate in a manner that istriggered to generate information as digital information that may be a record, which may be suitable for storage (e.g., temporarily, permanently, etc.) in a data storage device. As an example, a framework may operate in a manner whereby records are generated during execution of a simulator where the records may provide for auditing, re-running the simulator, assessment, quality control, etc.

[0114] While APIs are mentioned, one or more other technologies, techniques, etc., may be utilized that may provide for interactions between machines. As an example, a framework may include a simulator that may be a modified version of a commercial simulator in that the simulator may be controlled using information generated by one or more ML models. For example, modification may include insertion of hooks, calls, etc. As an example, an integrated approach may be utilized where, for example, a simulation framework is modified to include executable code for generating information using one or more ML models, which may be local, remote, local and remote, etc. As an example, one or more libraries may be utilized in a simulator control scheme. As an example, one or more background processes may be executed in a simulator control scheme. For example, consider an approach where training of one or more ML models may be performed as one or more background processes where, for example, training may improve ML model performance (e.g., timeliness, accuracy, resource demands, etc.). As an example, a system may provide for receipt of feedback during execution of a simulator whereby such feedback may be utilized to improve the system (e.g., via ML model selection, training, testing, etc.).

[0115] As an example, where sensor-based data as to salt may be available or become available, as explained, an architecture of a framework may be adaptable, for example, to select and utilize, or not, one or more ML models for handling salt. As explained, a salt metric may be compared to a threshold to determine whether salt is to be considered or not. As an example, a controller may be adaptable automatically responsive to receipt of data concerning salt. For example, a controller may determine that a salt concentration is above a threshold and cause the controller and / or one or more components operable coupled thereto to automatically adapt to handle salt such that controller may be improved and provide for improved control of one or more processes (e.g., via control of one or more pieces of field equipment). As an example, a controller may provide forautomatic selection of a particular EOS. For example, consider a controller that may compare predictions and sensor-based data where, if error is above a threshold, the controller may automatically adapt by selecting a different EOS and / or by running more than one EOS, for example, in parallel or series. In such an example, a controller may provide for determining which EOS may result in improved performance of the controller and, hence, improved control of a process.

[0116] FIG. 12 shows an example of a framework system 1200 that can include one or more frameworks 1210 that can implement one or more physicsbased components 1230 and one or more ML-based components 1250. As explained, a physics-based component 1230 may implement an EOS. For example, consider an EOS that may provide for relatively robust and acceptably accurate computation of values for fugacity coefficients, as may be associated with gas phase dynamics; whereas, for one or more other dynamics, one or more ML-based components may be utilized. As explained, an ML-based component may be tailored to a specific task, which may, for example, provide for acceptable accurate generation of output over a desired range of temperatures and a desired range of pressures.

[0117] As an example, the framework system 1200 may be utilized for one or more types of simulation workflows. For example, consider a reservoir simulation workflow, a process simulation workflow, etc. As an example, a simulator may perform a simulation in an iterative manner where convergence is aimed to be achieved after a number of iterations to arrive at a solution (e.g., simulation results). As an example, a method may include controlling a simulator based at least in part on one or more convergence criteria, which may include one or more error criteria, number of iterations criteria, etc. (e.g., consider use of an iterative techniques such as the Newton-Raphson technique, etc.). As an example, where a convergence issue may be detected, a framework may provide for controlling a simulator, for example, to cause the simulator to utilize a different technique or techniques for one or more computations. In such an example, the framework may cause the simulator to speed up, slow down, etc., depending on the type of issue detected. For example, if convergence is taking too many iterations to reach a solution for a system of equations, an inaccuracy may exist in one or more computations such that a framework may call for implementation of a more accurate computation technique,which may introduce additional demands for time and / or computational resources; whereas, if convergence is occurring quickly to an accurate solution, a framework may provide for assessing the circumstances (e.g., computational resources, size of problem, physical ranges of parameters, etc.) and grading a technique or techniques with respect to the circumstances. In such an approach, a framework may provide for learning when to adjust, change, etc., one or more techniques utilized by one or more simulators.

[0118] As an example, the framework system 1200 may provide for controlling a simulator to perform a simulation. As an example, the framework system 1200 may provide for controlling multiple simulators to perform multiple simulations. For example, consider a scenario where simulators may be executed at least in part in parallel, which may provide for generation of simulation results that may be statistically and / or otherwise assessed for improving one or more workflows (e.g., for planning, controlling, etc.).

[0119] As to types of machine learning (ML) models, consider one or more of a support vector machine (SVM) model, a k-nearest neighbors (KNN) model, an ensemble classifier model, a neural network (NN) model, etc. As an example, a machine learning model can be a deep learning model (e.g., deep Boltzmann machine, deep belief network, convolutional neural network, stacked auto-encoder, etc.), an ensemble model (e.g., random forest, gradient boosting machine, bootstrapped aggregation, AdaBoost, stacked generalization, gradient boosted regression tree, etc.), a neural network model (e.g., radial basis function network, perceptron, back-propagation, Hopfield network, etc.), a regularization model (e.g., ridge regression, least absolute shrinkage and selection operator, elastic net, least angle regression), a rule system model (e.g., cubist, one rule, zero rule, repeated incremental pruning to produce error reduction), a regression model (e.g., linear regression, ordinary least squares regression, stepwise regression, multivariate adaptive regression splines, locally estimated scatterplot smoothing, logistic regression, etc.), a Bayesian model (e.g., naive Bayes, average on-dependence estimators, Bayesian belief network, Gaussian naive Bayes, multinomial naive Bayes, Bayesian network), a decision tree model (e.g., classification and regression tree, iterative dichotomiser 3, C4.5, C5.0, chi-squared automatic interaction detection, decision stump, conditional decision tree, M5), a dimensionality reductionmodel (e.g., principal component analysis, partial least squares regression, Sammon mapping, multidimensional scaling, projection pursuit, principal component regression, partial least squares discriminant analysis, mixture discriminant analysis, quadratic discriminant analysis, regularized discriminant analysis, flexible discriminant analysis, linear discriminant analysis, etc.), an instance model (e.g., k- nearest neighbor, learning vector quantization, self-organizing map, locally weighted learning, etc.), a clustering model (e.g., k-means, k-medians, expectation maximization, hierarchical clustering, etc.), etc.

[0120] As an example, a machine model may be built using a computational framework with a library, a toolbox, etc., such as, for example, those of the MATLAB framework (MathWorks, Inc., Natick, Massachusetts). The MATLAB framework includes a toolbox that provides supervised and unsupervised machine learning algorithms, including support vector machines (SVMs), boosted and bagged decision trees, k-nearest neighbor (KNN), k-means, k-medoids, hierarchical clustering, Gaussian mixture models, and hidden Markov models. Another MATLAB framework toolbox is the Deep Learning Toolbox (DLT), which provides a framework for designing and implementing deep neural networks with algorithms, pretrained models, and apps. The DLT provides convolutional neural networks (ConvNets, CNNs) and long short-term memory (LSTM) networks to perform classification and regression on image, time-series, and text data. The DLT includes features to build network architectures such as generative adversarial networks (GANs) and Siamese networks using custom training loops, shared weights, and automatic differentiation. The DLT provides for model exchange various other frameworks.

[0121] As an example, the TENSORFLOW framework (Google LLC, Mountain View, CA) may be implemented, which is an open-source software library for dataflow programming that includes a symbolic math library, which can be implemented for machine learning applications that can include neural networks. As an example, the CAFFE framework may be implemented, which is a DL framework developed by Berkeley Al Research (BAIR) (University of California, Berkeley, California). As another example, consider the SCIKIT platform (e.g., scikit-learn), which utilizes the PYTHON programming language. As an example, a framework such as the APOLLO Al framework may be utilized (APOLLO. Al GmbH, Germany).As an example, a framework such as the PYTORCH framework may be utilized (Facebook Al Research Lab (FAIR), Facebook, Inc., Menlo Park, California).

[0122] As an example, a training method can include various actions that can operate on a dataset to train a ML model. As an example, a dataset can be split into training data and test data where test data can provide for evaluation. A method can include cross-validation of parameters and best parameters, which can be provided for model training.

[0123] The TENSORFLOW framework can run on multiple CPUs and GPUs (with optional CUDA (NVIDIA Corp., Santa Clara, California) and SYCL (The Khronos Group Inc., Beaverton, Oregon) extensions for general-purpose computing on graphics processing units (GPUs)). TENSORFLOW is available on 64-bit LINUX, MACOS (Apple Inc., Cupertino, California), WINDOWS (Microsoft Corp., Redmond, Washington), and mobile computing platforms including ANDROID (Google LLC, Mountain View, California) and IOS (Apple Inc.) operating system-based platforms.

[0124] TENSORFLOW computations can be expressed as stateful dataflow graphs; noting that the name TENSORFLOW derives from the operations that such neural networks perform on multidimensional data arrays. Such arrays can be referred to as “tensors”.

[0125] As an example, one or more features of the KERAS library may be utilized. The KERAS library is an open-source library that provides a PYTHON interface for artificial neural networks (ANNs). The KERAS library can act as an interface for the TENSORFLOW library.

[0126] As an example, a device may utilize TENSORFLOW LITE (TFL) or another type of lightweight framework. TFL is a set of tools that enables on-device machine learning where models may run on mobile, embedded, and loT devices. TFL is optimized for on-device machine learning, by addressing latency (no roundtrip to a server), privacy (no personal data leaves the device), connectivity (Internet connectivity is demanded), size (reduced model and binary size) and power consumption (e.g., efficient inference and a lack of network connections). TFL includes multiple platform support, covering ANDROID and iOS devices, embedded LINUX, and microcontrollers. TLF provides diverse language support, which includes JAVA, SWIFT, Objective-C, C++, and PYTHON. TFL provides high performance, with hardware acceleration and model optimization. As an example,one or more machine learning tasks may include, for example, classification, regression, object detection, pose estimation, question answering, text classification, etc., on one or more of multiple platforms.

[0127] FIG. 13 shows an example of a method 1300 and an example of a system 1390. As shown, the method 1300 can include a control block 1310 for controlling a simulator to perform a simulation as to component interactions for gas and liquid components in a subsurface region; a computation block 1320 for, during the simulation, computing fugacity values using an equation of state, computing an equilibrium constant value using a first machine learning model, computing a Henry’s law constant value using a second machine learning model, and computing component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constant value; and a generation block 1330 for generating simulation results for the subsurface region based at least in part on the component mole fraction values.

[0128] The method 1300 is shown in FIG. 13 in association with various computer-readable media (CRM) blocks 1311 , 1321 , and 1331. Such blocks generally include instructions suitable for execution by one or more processors (or processor cores) to instruct a computing device or system to perform one or more actions. While various blocks are shown, a single medium may be configured with instructions to allow for, at least in part, performance of various actions of the method 1300. As an example, a computer-readable medium (CRM) may be a computer-readable storage medium that is non-transitory and that is not a carrier wave. As an example, one or more of the blocks 1311 , 1321 , and 1331 may be in the form processor-executable instructions.

[0129] In the example of FIG. 13, the system 1390 includes one or more information storage devices 1391 , one or more computers 1392, one or more networks 1395 and instructions 1396. As to the one or more computers 1392, each computer may include one or more processors (e.g., or processing cores) 1393 and memory 1394 for storing the instructions 1396, for example, executable by at least one of the one or more processors 1393 (see, e.g., the blocks 1311 , 1321 , and 1331 ). As an example, a computer may include one or more network interfaces (e.g., wired or wireless), one or more graphics cards, a display interface (e.g., wired or wireless), etc.

[0130] As an example, a method can include performing a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, computing fugacity values using an equation of state, computing an equilibrium constant value using a first machine learning model, computing a Henry’s law constant value using a second machine learning model, and computing component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constant value; and generating simulation results for the subsurface region based at least in part on the component mole fraction values. In such an example, the components may include hydrogen and water.

[0131] As an example, an equation of state may depend on an input temperature and an input pressure.

[0132] As an example, a method may employ multiple machine learning models where, for example, a first machine learning model generates an equilibrium constant value based on an input temperature and an input pressure and, for example, where a second machine learning model generates a Henry’s law constant value based on an input temperature and an input pressure.

[0133] As an example, component mole fraction values in a gas phase and in a liquid phase may correspond to a grid cell of a grid cell simulation model of a subsurface region. In such an example, the grid cell of the grid cell simulation model of the subsurface region may include a temperature and a pressure.

[0134] As an example, a simulation may include a hydrogen storage simulation for storage of hydrogen in a subsurface region and / or a simulation may include a hydrogen production simulation for production of hydrogen from a storage of hydrogen in a subsurface region.

[0135] As an example, a method may include using multiple machine learning models where a first machine learning model may be or include a neural network model and a second machine learning model may be or include a neural network model.

[0136] As an example, a method may include a machine learning model that generates an activity coefficient value. For example, consider a machine learning model that generates the activity coefficient value based in part on an input temperature. In such an example, components may include salt where, for example,the machine learning model generates the activity coefficient value based in part on a molality of salt in the liquid phase. As an example, such a machine learning model may be one of three or more machine learning models. For example, consider a method that may implement a first machine learning model, a second machine learning model, and a third machine learning model.

[0137] As an example, a method can include controlling a field operation based at least in part on simulation results. In such an example, the field operation may include one or more of an injection operation that injects at least one of a number of components into a subsurface region and a production operation that produces at least one of a number of components from a subsurface region.

[0138] As an example, components may include one or more of water, hydrogen, carbon dioxide, methane, and hydrogen sulfide.

[0139] As an example, a system can include a processor; a memory operatively coupled to the processor; processor-executable instructions stored in the memory and executable to instruct the system to: perform a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, compute fugacity values using an equation of state, compute an equilibrium constant value using a first machine learning model, compute a Henry’s law constant value using a second machine learning model, and compute component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constant value; and generate simulation results for the subsurface region based at least in part on the component mole fraction values.

[0140] As an example, one or more computer-readable storage media may include processor-executable instructions executable by a system to instruct the system to: perform a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, compute fugacity values using an equation of state, compute an equilibrium constant value using a first machine learning model, compute a Henry’s law constant value using a second machine learning model, and compute component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constant value; and generate simulation results for the subsurface region based at least in part on the component mole fraction values.

[0141] As an example, a computer program product can include one or more computer-readable storage media that can include processor-executable instructions to instruct a computing system to perform one or more methods and / or one or more portions of a method.

[0142] In some embodiments, a method or methods may be executed by a computing system. FIG. 14 shows an example of a system 1400 that can include one or more computing systems 1401 -1 , 1401 -2, 1401 -3 and 1401 -4, which may be operatively coupled via one or more networks 1409, which may include wired and / or wireless networks.

[0143] As an example, a system can include an individual computer system or an arrangement of distributed computer systems. In the example of FIG. 14, the computer system 1401 -1 can include one or more modules 1402, which may be or include processor-executable instructions, for example, executable to perform various tasks (e.g., receiving information, requesting information, processing information, simulation, outputting information, etc.).

[0144] As an example, a module may be executed independently, or in coordination with, one or more processors 1404, which is (or are) operatively coupled to one or more storage media 1406 (e.g., via wire, wirelessly, etc.). As an example, one or more of the one or more processors 1404 can be operatively coupled to at least one of one or more network interfaces 1407; noting that one or more other components 1408 may also be included. In such an example, the computer system 1401 -1 can transmit and / or receive information, for example, via the one or more networks 1409 (e.g., consider one or more of the Internet, a private network, a cellular network, a satellite network, etc.).

[0145] As an example, the computer system 1401-1 may receive from and / or transmit information to one or more other devices, which may be or include, for example, one or more of the computer systems 1401 -2, etc. A device may be located in a physical location that differs from that of the computer system 1401-1 . As an example, a location may be, for example, a processing facility location, a data center location (e.g., server farm, etc.), a rig location, a wellsite location, a downhole location, etc.

[0146] As an example, a processor may be or include a microprocessor, microcontroller, processor module or subsystem, programmable integrated circuit, programmable gate array, or another control or computing device.

[0147] As an example, the storage media 1406 may be implemented as one or more computer-readable or machine-readable storage media. As an example, storage may be distributed within and / or across multiple internal and / or external enclosures of a computing system and / or additional computing systems.

[0148] As an example, a storage medium or storage media may include one or more different forms of memory including semiconductor memory devices such as dynamic or static random access memories (DRAMs or SRAMs), erasable and programmable read-only memories (EPROMs), electrically erasable and programmable read-only memories (EEPROMs) and flash memories, magnetic disks such as fixed, floppy and removable disks, other magnetic media including tape, optical media such as compact disks (CDs) or digital video disks (DVDs), BLUERAY disks, or other types of optical storage, or other types of storage devices.

[0149] As an example, a storage medium or media may be located in a machine running machine-readable instructions, or located at a remote site from which machine-readable instructions may be downloaded over a network for execution. As an example, various components of a system such as, for example, a computer system, may be implemented in hardware, software, or a combination of both hardware and software (e.g., including firmware), including one or more signal processing and / or application specific integrated circuits.

[0150] As an example, a system may include a processing apparatus that may be or include a general-purpose processors or application specific chips (e.g., or chipsets), such as ASICs, FPGAs, PLDs, or other appropriate devices.

[0151] As an example, a device may be a mobile device that includes one or more network interfaces for communication of information. For example, a mobile device may include a wireless network interface (e.g., operable via IEEE 802.11 , ETSI GSM, BLUETOOTH, satellite, etc.). As an example, a mobile device may include components such as a main processor, memory, a display, display graphics circuitry (e.g., optionally including touch and gesture circuitry), a SIM slot, audio / video circuitry, motion processing circuitry (e.g., accelerometer, gyroscope), wireless LAN circuitry, smart card circuitry, transmitter circuitry, GPS circuitry, and abattery. As an example, a mobile device may be configured as a cell phone, a tablet, etc. As an example, a method may be implemented (e.g., wholly or in part) using a mobile device. As an example, a system may include one or more mobile devices.

[0152] As an example, a system may be a distributed environment, for example, a so-called “cloud” environment where various devices, components, etc. interact for purposes of data storage, communications, computing, etc. As an example, a device or a system may include one or more components for communication of information via one or more of the Internet (e.g., where communication occurs via one or more Internet protocols), a cellular network, a satellite network, etc. As an example, a method may be implemented in a distributed environment (e.g., wholly or in part as a cloud-based service).

[0153] As an example, information may be input from a display (e.g., consider a touchscreen), output to a display or both. As an example, information may be output to a projector, a laser device, a printer, etc. such that the information may be viewed. As an example, information may be output stereographically or holographically. As to a printer, consider a 2D or a 3D printer. As an example, a 3D printer may include one or more substances that can be output to construct a 3D object. For example, data may be provided to a 3D printer to construct a 3D representation of a subterranean formation. As an example, layers may be constructed in 3D (e.g., horizons, etc.), geobodies constructed in 3D, etc. As an example, holes, fractures, etc., may be constructed in 3D (e.g., as positive structures, as negative structures, etc.).

[0154] Although only a few example embodiments have been described in detail above, those skilled in the art will readily appreciate that many modifications are possible in the example embodiments. Accordingly, all such modifications are intended to be included within the scope of this disclosure as defined in the following claims. In the claims, means-plus-function clauses are intended to cover the structures described herein as performing the recited function and not only structural equivalents, but also equivalent structures. Thus, although a nail and a screw may not be structural equivalents in that a nail employs a cylindrical surface to secure wooden parts together, whereas a screw employs a helical surface, in theenvironment of fastening wooden parts, a nail and a screw may be equivalent structures.

[0155] Bibliography (documents incorporated by reference herein)1 . Sainz-Garcia A, Abarca E, Rubi V, Grandia F. Assessment of feasible strategies for seasonal underground hydrogen storage in a saline aquifer. Int J Hydrogen Energy 2017; 42: 16657-66. https: / / doi.Org / 10.1016 / j . ijhydene.2017.05.076.2. Heinemann N, Alcalde J, Miocic JM, Hangx SJT, Kallmeyer J, Ostertag- Henning C, et al. Enabling large-scale hydrogen storage in porous media — The scientific challenges. Energy Environ Sci 2021 ; 14: 853-64. https: / / doi.Org / 10.1039 / d0ee03536j.3. Li D, Beyer C, Bauer S. A unified phase equilibrium model for hydrogen solubility and solution density. Int J Hydrogen Energy 2018; 43(1 ): 512-29.4. Chabab S, Theveneau P, Coquelet C, Corvisier J, Paricaud P. Measurements and predictive models of high-P H2 solubility in brine (H2O-NaCI) for underground hydrogen storage application. Int J Hydrogen Energy 2020; 45(56): 32206-20.5. Rahbari A, Brenkman J, Hens R, Ramdin M, Van Den Broeke LJ, Schoon R, et al. Solubility of water in hydrogen at high pressures: A molecular simulation study. J Chem Eng Data 2019; 64(9): 4103-15.6. Lopez-Lazaro C, Bachaud P, Moretti I, Fernando N. Predicting the phase behavior of hydrogen in NaCI brines by molecular simulation for geological applications. BSGF-Earth Sci Bull 2019; 190(1 ): 7. https: / / doi.Org / 10.1051 / bsgf / 2019008.7. Zhu Z, Cao Y, Zheng Z, Chen D. An accurate model for estimating H2 solubility in pure water and aqueous NaCI solutions. Energies 2022; 15: 5021-36. https: / / doi.Org / 10.3390 / en151450218. Kontogeorgis GM, Yakoumis IV, Meijer H, Hendriks EM, Moorwood T. Multicomponent phase equilibrium calculations for water-methanol-alkanemixtures. Fluid Phase Equilibria 1999; 158(160): 201-9.9. Peng DY, Robinson DB. A new two-constant equation of state. Ind Eng Chem Fundam 1976; 15: 59-64. https: / / doi.org / 10.1021 / i160057a011.10. Zhao M, Okuno R. A proxy Peng-Robinson EOS for efficient modeling of phase behavior. SPE Reservoir Simulation Conference 2021 ; October 4-6, Galveston, Texas, USA.11 . Ansari S, Safaei-Farouji M, Atashrouz S, Abedi A, Hemmati-Sarapardeh A, Mohaddespour A. Prediction of hydrogen solubility in aqueous solu-tions: Comparison of equations of state and advanced machine learning-metaheuristic approaches. Int J Hydrogen Energy, 2022; 47(89): 37724-41. https: / / doi.Org / 10.1016 / j . ijhydene.2022.08.288.12. Karniadakis GE, Kevrekidis IG, Lu L. Physics-informed machine learning. Nat Rev Phys 2021 ; 3: 422-440. https: / / doi.org / 10.1038 / s42254-021-00314-5.13. Gillespie PC, Wilson GM. Vapor-liquid equilibrium data on water-substitute gas components: N2-H2O, H2-H2O, CO-H2O, H2-CO-H2O, and H2S-H2O. 1980. United States, https: / / doi.org / 10.2172 / 6782591.14. Wiebe R, Gaddy VL. The solubility of hydrogen in water at 0, 50, 75 and 100 from 25 to 1000 atmospheres. J Am Chem Soc 1934; 56: 76e9. https: / / doi.Org / 10.1021 / ja01316a022.15. Ugrozov W. Equilibrium compositions of vapor-gas mixtures over solutions. Russ J Phys Chem 1996; 70(7): 1240-1.16. Maslennikova VY, Goryunova NP, Subbotina LA, Tsiklis DS. The solubility of water in compressed hydrogen. Russ J Phys Chem B 1976; 50: 240-3.17. Bartlett EP. The concentration of water vapor in compressed hydrogen, nitrogen and a mixture of these gases in the presence of condensed water. J Am Chem Soc 1927; 49(1 ): 65-78.18. Devaney W, Berryman JM, Kao PL, Eakin B. High temperature VLE measurements for substitute gas components. Gas Proc Assoc 1978.19. Kling G, Maurer G. The solubility of hydrogen in water and in 2-aminoethanol at temperatures between 323 K and 423 K and pressures up to 16 MPa. J Chem Thermo 1991 ; 23(6): 531-41.20. Jung J, Knacke O, Neuschutz D. Ldslichkeit von kohlenmonoxid und was- serstoff in wasser bis 300°C. Chemie Ingenieur Technik 1971 ; 43(3): 112-6.21 . Ipatev V, Teodorovich V. Equilibrium compositions of vapor-gas mixtures over solutions. Zh Obshch Khim 1934; 4: 395-9.

Claims

CLAIMSWhat is claimed is:1 . A method comprising: controlling a simulator to perform a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, computing fugacity values using an equation of state, computing an equilibrium constant value using a first machine learning model, computing a Henry’s law constant value using a second machine learning model, and computing component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constant value; and generating simulation results for the subsurface region based at least in part on the component mole fraction values.

2. The method of claim 1 , wherein the components comprise hydrogen and water.

3. The method of claim 1 , wherein the equation of state depends on an input temperature and an input pressure.

4. The method of claim 1 , wherein the first machine learning model generates the equilibrium constant value based on an input temperature and an input pressure.

5. The method of claim 1 , wherein the second machine learning model generates the Henry’s law constant value based on an input temperature and an input pressure.

6. The method of claim 1 , wherein the component mole fraction values in the gas phase and in the liquid phase correspond to a grid cell of a grid cell simulation model of the subsurface region.

7. The method of claim 6, wherein the grid cell of the grid cell simulation model of the subsurface region comprises a temperature and a pressure.

8. The method of claim 1 , wherein the simulation comprises a hydrogen storage simulation for storage of hydrogen in the subsurface region.

9. The method of claim 1 , wherein the simulation comprises a hydrogen production simulation for production of hydrogen from a storage of hydrogen in the subsurface region.

10. The method of claim 1 , wherein the first machine learning model comprises a neural network model.11 . The method of claim 1 , wherein the second machine learning model comprises a neural network model.

12. The method of claim 1 , comprising a third machine learning model that generates an activity coefficient value.

13. The method of claim 12, wherein the third machine learning model generates the activity coefficient value based in part on an input temperature.

14. The method of claim 13, wherein the components comprise salt.

15. The method of claim 14, wherein the third machine learning model generates the activity coefficient value based in part on a molality of salt in the liquid phase.

16. The method of claim 1 , comprising controlling a field operation based at least in part on the simulation results.

17. The method of claim 16, wherein the field operation comprises one or more of an injection operation that injects at least one of the components into the subsurfaceregion and a production operation that produces at least one of the components from the subsurface region.

18. The method of claim 1 , wherein the components comprise one or more of water, hydrogen, carbon dioxide, methane, and hydrogen sulfide.

19. A system comprising: a processor; a memory operatively coupled to the processor; processor-executable instructions stored in the memory and executable to instruct the system to: perform a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, compute fugacity values using an equation of state, compute an equilibrium constant value using a first machine learning model, compute a Henry’s law constant value using a second machine learning model, and compute component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constant value; and generate simulation results for the subsurface region based at least in part on the component mole fraction values.

20. One or more computer-readable storage media comprising processor-executable instructions executable by a system to instruct the system to: perform a simulation as to component interactions for gas and liquid components in a subsurface region; during the simulation, compute fugacity values using an equation of state, compute an equilibrium constant value using a first machine learning model,compute a Henry’s law constant value using a second machine learning model, and compute component mole fraction values in a gas phase and in a liquid phase based on the fugacity values, the equilibrium constant value, and the Henry’s law constant value; and generate simulation results for the subsurface region based at least in part on the component mole fraction values.

Images