WO2025229366A1

WO2025229366A1 - A method for modelling a fractured geological reservoir comprising fault arrays or fracture corridors

Info

Publication number: WO2025229366A1
Application number: PCT/IB2024/000211
Authority: WO
Inventors: Cédric GAL; LEONFORTE (GARI), Emmanuelle; Gérard MASSONNAT
Original assignee: TotalEnergies Onetech SAS
Current assignee: TotalEnergies Onetech SAS
Priority date: 2024-04-30
Filing date: 2024-04-30
Publication date: 2025-11-06
Anticipated expiration: 2026-10-30

Abstract

It is disclosed a computer-implemented method for modelling a reservoir comprising fault arrays or fracture corridors, comprising: - obtaining an initial model of the reservoir (100) and ground data (200), - Implementing (300) at least one set of fault arrays or fracture corridors generation in the initial model, comprising receiving input parameters (310), and, based on the received ground data and input parameters: o generating a plurality of array or corridor seeds (321) in the initial model, and, o For at least one seed, ▪ Generating (322), from the seed, a volume of the fault array or fracture corridor, ▪ Generating (323), within said volume, a plurality of fault segments seeds, and ▪ Generating (324) fault segments from at least part of the fault segments seeds.

Description

A METHOD FOR MODELLING A FRACTURED GEOLOGICAL RESERVOIR COMPRISING FAULT ARRAYS OR FRACTURE CORRIDORS

TECHNICAL FIELD

The present disclosures relates to a computer-implemented method for modelling a geological reservoir incorporating fractures, in particular fault arrays or fracture corridors. The disclosure finds notable applications in the fields of hydrocarbon production, Carbon Capture and Storage (CCS), or water resource management.

TECHNICAL BACKGROUND

In geology, there is an ongoing need of proposing computer modelling methods that enable accurately representing real geological reservoirs. Such models are indeed used to simulate various scenarios and predict how a reservoir might behave under different conditions.

In the field of hydrocarbon production, a good understanding of the actual geometry and petrophysical properties of a reservoir are essential for forecasting fluid flow patterns, and determining oil and gas recovery strategies. In the field of carbon capture and storage, it is equally important to rely on accurate representations of the reservoir in order to forecast the storage potential of a reservoir, the most appropriate locations of injection wells, and anticipate risks of leakage. Also in the field of water resource management, relying on an accurate reservoir representation enables simulating the movement and behavior of groundwater in aquifers, enabling prediction of water availability, as well as potential flooding or pollution risks.

One element that can strongly influence the behavior of a reservoir relates to the existence and disposition of fractures within the reservoir. Indeed, fractures creates local heterogeneities in the behavior of a reservoir, in particular with respect to permeability and hence to fluid flows within the reservoir.

Furthermore, the generic term of “fractures” actually encompasses a large variety of geological structures, exhibiting different geometries, respective chronologies of formation and hence different impacts as to the petrophysical properties of a geological reservoir. Geologists thus have established a typology of fractures, according to features such as the mode of deformation of the fractures, or their specific chronology of formation. One may for instance refer to [Peacock, 2016]. In the field of hydrocarbon production, an accurate representation of the disposition of fractures is essential for forecasting production, whereas in the field of CCS, fractures may cause leaks and thus adversely affect the efficiency of carbon storage. There is therefore a need for a good knowledge of fractures present within a reservoir in order to accurately represent said reservoir and use it for the above-mentioned applications.

However, the available ground data is too scarce to accurately represent the fractures within a reservoir. Regarding ground data that is obtained through campaigns of seismic reflections, it only enables detecting major fractures, but fails to provide any information on fractures of lesser size. As to ground data that is obtained by drilling exploration wells and analyzing the wells, it provides detailed but local information, and fails to provide an extensive hindsight about the number, disposition and relationship between fractures within all the considered domain.

In this perspective, computer modelling methods enabling the generation of fractures have been developed. Simulation of fractures is commonly achieved by generating a Discrete Fracture Network (DFN), which is a numerical representation of fractures, in which fractures are represented as discrete features associated with specific properties, including orientation, length, aperture and connectivity.

In particular, it is known from [Libby, 2019], an algorithm for modelling fractures, according to which fractures are grown according to a sequence of fracture sets. Within each set, fractures grow from a seed point stochastically located in space, until it reaches its target size, terminates against a previous fracture, or until a target fracture intensity is reached. Exclusion zones are defined around each fracture that prevent nucleation of new fractures of the same set within the exclusion zones. Moreover, for a considered fracturing set, a user may configure the geometrical parameters (size, orientation) of the fractures, as well as rules of interaction between fractures of a same set. This algorithm thus enables modelling fractures with varying orientations, sizes and types of intersection.

However, this algorithm still does not enable accurate representation of the diversity of fracture types within a reservoir, as well as the interactions between them. First, in this algorithm, all fractures are generated stochastically based on density data, and no fracture which location is based on hard data may be integrated. Second, despite changes in sizes and orientations of the generated fractures, all fractures are generated according to the same process, and the algorithm does not manage different rules of intersection between fractures of different fracturing sets. To the contrary, as all fractures and generated and represented according to the same process, the rules of intersection between fractures are also always the same. Therefore the resulting network of fractures does not accurately reflect reality, and hence the resulting upscaling and/or fluid flow simulations that will be performed might be flawed.

SUMMARY OF THE INVENTION

An aim of the present disclosure is to improve the situation.

In particular, an aim of the present disclosure is to provide a method for modelling a reservoir including fractures, which more accurately reflects the disposition of the fractures of the reservoir, the actual geological phenomena that have led to the formation of the fractures, and the impact of these fractures on the petrophysical properties and behavior of the geological reservoir, when said fractures are fault arrays or fracture corridors.

Another aim of the present disclosure is to provide a method enabling accurately modelling the formation of fault arrays or fracture corridors. Another aim of the present disclosure is to provide a method enabling modelling a chronological sequence of fracturing sets, where a subsequent fracturing set takes into account the fractures generated during an earlier fracturing set.

To this end, it is disclosed a computer-implemented method for modelling a reservoir comprising fault arrays or fracture corridors, comprising: obtaining an initial model of the reservoir and ground data, Implementing at least one set of fault arrays or fracture corridors generation in the initial model, comprising receiving input parameters, and, based on the received ground data and input parameters: o generating a plurality of array or corridor seeds in the initial model, and, o For at least one seed, ■ Generating, from the seed, a volume of the fault array or fracture corridor,

■ Generating, within said volume, a plurality of fault segments seeds, and

■ Generating fault segments from at least part of the fault segments seeds.

In embodiments, the ground data comprises at least one of: a localization of fractures, a density of fractures corridors, fault arrays, or of fractures within a fracture corridor or a fault array, geometric parameters of fracture corridors, fault arrays or fractures within a fracture corridor or a fault array, a chronology of the formation of fractures.

In embodiments, generating the array or corridor seeds is performed by implementation of a probabilistic process based on a predetermined density.

In embodiments, generating fault segments seeds is performed by implementation of a stochastic process based on a predetermined density.

In embodiments, the method comprises implementing at least two sets of fault arrays generation, and the method comprises developing fault segments from a segment seed when a line joining the fault segment seed and the seed of the array to which belongs the fault segment-does not cross a fault segment of a fault array of a previous set.

In embodiments, the generated fault arrays are selected among a plurality of types of fault arrays, where each type of fault arrays is defined by:

A number of fault segments families,

Geometric parameters associated with each fault segment families, Whether or not a fault segment is associated to a stress shadow zone, and When the number of fault segments families is greater than 1 : o intersection rules between fault segments of a first family and fault segments of other families, and o order of generation of fault segments of each fault segment family.

In embodiments, the method comprises at least one set of generation of relay faults, wherein relay faults comprise two fault segment families associated to respective geometric parameters, fault segments of a first family extending in a first direction, and fault segments of the second family extending in a second direction, transverse to the first direction, and developing fault segments from the fault segments seeds comprises:

For each seed, when the seed does not belong to a stress shadow zone of a previously generated fault segment, generating a fault segment of the first family, and

For at least one pair of fault segments of the first family, when the two fault segments overlap at least in part and a distance between the two fault segments is below a determined threshold, generating a fault segment extending between the two fault segments of the first family and abutting against them.

In embodiments, the method comprises at least one set of generation of anastomosing faults, wherein anastomosing faults comprise two fault segment families associated to respective orientation parameters, fault segments of a first family extending in a first direction, and fault segments of the second family extending in a second direction, transverse to the first direction, and developing fault segments from the fault segments seeds comprise generating alternatively a fault segment of the first family and a first segment of the second family, wherein, in case of intersection between a currently generated fault segment and a previously generated fault segment, the currently generated fault segment abuts against the previously generated fault segment regardless the family of the previous fault segment.

In embodiments, the method comprises at least one set of generation of en-echelon faults, wherein the fault segments seeds are generated along a line extending parallel to the main direction of the array, and all fault segments generated from the fault segments seeds extend parallel to each other. In embodiments, the method further comprises a step of assigning at least one of an aperture or permeability value to the generated fault segments.

In embodiments, the method further comprising a step of upscaling the obtained reservoir model or the obtained fault arrays.

According to another object, it is disclosed a computer program product comprising code instructions for implementing the method according to the description above, when it is executed by a computer.

According to another object, it is disclosed a non-transitory computer-readable storage having stored thereon code instructions which, when executed by a computer, cause said computer to implement the method according to the description above.

According to another object; it is disclosed a computer, configured for implementing the method according to the description above.

The disclosed method enables modelling the formation of a geological reservoir including fracture corridors or fault arrays, while respecting the specificities of these geological objects in terms of formation, disposition and hence petrophysical behaviour.

The disclosed method enables managing rules of intersection between fault segments of a given set, i.e. of a given fault array of fracture corridor, while also taking into account previous generated fault arrays or fracture corridors in the formation of latter fault segments.

Thanks to this method, studies and exploitation of real geological reservoirs, with a perspective of exploiting the resources they contain or are able to contain (regarding the field of carbon storage for instance), may be performed more reliably.

BRIEF DESCRIPTION OF THE DRAWINGS

The present disclosure is illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings, in which like reference numerals reference to similar elements and in which:

Figures 1 a and 1 b are flow charts describing the main steps of a possible embodiments of a method for modelling a reservoir, Figures 2a to 2c represent various types of fault arrays including respectively relay faults, anastomosing faults and en-echelon faults.

Figure 3 schematically represents an example of a geological calendar, Figure 4 schematically represents geometrical parameters of a fracture, Figure 5 schematically represents a fault array generation process according to a possible embodiment,

Figure 6 schematically represents a rule for managing intersection between fault arrays of successive fracturing sets,

Figures 7a to 7c represents how are modelled relay faults, anastomosing faults and en-echelon faults respectively within the present disclosure, Figure 8 is a possible embodiment of a device for implementing the method according to the disclosure

Figures 9a and 9b schematically represent the computation of pseudoequivalent fractures to represent a plurality of fault segments of a fracture corridor.

DESCRIPTION OF AT LEAST ONE EMBODIMENT

A method for modelling reservoirs incorporating fractures according to embodiments of the present disclosure will now be described.

The method may be carried out by a computer system 1 , schematically represented in figure 8. In preferred embodiments, the computer system 1 comprises one or more processors 10 (which may belong to a same computer or to different computers) and storage means 20 (magnetic hard disk, optical disk, electronic memory, or any computer readable storage medium) in which a computer program product is stored, in the form of a set of program-code instructions to be executed in order to implement all or part of the steps of the method. Alternatively, or in combination thereof, the computer system can comprise one or more programmable logic circuits (FPGA, PLD, etc.), and/or one or more specialized integrated circuits (ASIC), etc., adapted for implementing all or part of said steps of the processing method. In other words, the computer system comprises a set of means configured by software (specific computer program product) and/or by hardware (processor, FPGA, PLD, ASIC, etc.) to implement the steps of the method. The computer system may comprise an input interface 30 for the reception of input data, in particular ground data regarding a reservoir that is to be modelled by application of the method. The input interface may be an interface with a storage device storing the input data, or a communication interface for connecting the computer system to a telecommunication network enabling downloading of said input data therefrom. The computer system may also comprise a display 40 for displaying three-dimensional representations of the obtained model of a reservoir, and/or two- dimensional representations, which may be sectional view of the model, and/or any relevant information relative to the obtained model.

A usual categorization of fractures is based on presence or absence of relative movement between the rocks on either side of the fracture. Accordingly, a first category of fractures are joints, which refer to a fracture in a rock mass where there has been no significant movement or displacement of the rocks on either side of the fracture. Joints grow perpendicular to stratification. A second category of fractures are faults, which refer to fractures in a rock mass due to a relative displacement of rocks on either side of the fracture, where this displacement can be for instance: a horizontal sliding of the rock masses on both sides of the fracture, a vertical sliding, resulting from either a compressional or extensional relative movement of the rock masses on both sides of the fracture.

The method disclosed below enables modelling a plurality of pre-determined types of fractures, where each fracture type is associated with a respective, specific generation process that is designed to accurately reflect the geometry and geological behavior of corresponding real fractures.

The plurality of pre-determined types of fractures may include: fault arrays, which refer to a collection of faults that occur in a particular region, and which may be related to regional tectonic forces, reflecting larger-scale deformation of the Earth’s crust. These faults may be related in terms of their orientation, arrangement, formation time or tectonic pattern. fractures clusters, which represent concentrations of fractures in a specific geological setting. These fractures may not necessarily involve significant displacement, and they can occur on various scales. The clusters may result from localized stress concentrations, variations in lithology or other factors influencing the mechanical behavior of the rocks. The fractures within a cluster may exhibit varying orientations and lengths, and they may not necessarily form a systematic pattern.

In embodiments, the method enables modelling various types of fault arrays, including the following three sub-types of fault arrays, examples of which are represented in figures 2a to 2c: en-echelon faults, shown in figure 2a, refer to a set of parallel faults that are arranged in a step-like pattern. These faults may not be perfectly aligned but are somewhat staggered like the arrangement of rungs on a ladder. En- echelon faults can appear in extensional tectonic settings. relay faults, shown in figure 2b, refer to a set of faults comprising short, linking segments between two larger fault segments, enabling transfer of displacement therebetween. anastomosing faults, shown in figure 2c, refer to a network of interconnected faults that merge and diverge, creating a complex pattern. Anastomosing faults are common in regions where there are multiple phases of deformation or in areas with heterogeneous rock properties.

In embodiments, regarding fracture clusters, the method enables modelling a specific type of fracture cluster which is fracture corridors. A fracture corridor refers to an elongated zone in which numerous faults and joints are aligned along a common trend. These corridors represent pathways where the Earth’s crust has experiences significant fracturing, often influenced by geological processes or stress fields.

Referring to figures 1 a and 1 b, the main steps of a method for modelling a geological reservoir including fractures corridors or fault arrays will now be disclosed. The method may in particular be used for modelling a real reservoir, i.e. a reservoir existing in real life, and for which ground data describing at least some geological and/or geometrical properties is available or can be acquired. The ground data can notably include, or be derived from, campaigns of seismic reflections, in-situ observations of geologists, satellite or aerial images, or wellbore data, including well logs, cores and plugs. The method thus enables forming more accurate and complete models of real reservoirs, for analyzing said models and using them for various applications, for instance for performing fluid-flow simulations, predicting reservoir production, determining location of injection and/or production wells, determining production scenarios, evaluating carbon storage capacities, determining fluid circulation or leakage within the reservoir, evaluating risks of pollution of ground water, etc.

The method comprises a step 100 of obtaining an initial model of the reservoir. The initial model of the reservoir is a numerical representation of the three-dimensional arrangement or rock structures forming the reservoir, which is defined at least by a three-dimensional volume V of determined dimensions.

In embodiments, the initial model of the reservoir further comprises a plurality of surfaces extending within said volume. The plurality of surfaces may include a plurality of stratigraphic surfaces, separating successive layers of rocks called beds. In embodiments, the initial model of the reservoir may comprise a plurality of distinct mechanical units, where each mechanical unit comprises a plurality of beds, and two successive beds, or two successive mechanical units are separated by a stratigraphic surface.

In what follows, a mechanical unit is defined as a geological formation comprising a plurality of beds, and whose upper and lower surfaces limit the vertical persistence of a given type of fracture. Rocks within a mechanical unit usually share similar properties, such as strength and deformation characteristics.

A bed refers to a layer of sedimentary rock, which results from the deposition of sediments over time. Each bed represents a single episode of sedimentation, often characterized by features such as grain size, mineral composition, level of sorting, or other sedimentary features.

At least one well may also be represented in the initial model of the reservoir, at a location corresponding to its actual location relative to the reservoir to be modelled. The well may in particular be an exploration well from which ground data has been acquired.

The method also comprises a step 200 of receiving ground data relative to the reservoir.

The ground data relative to the reservoir may comprise a chronology of the formation of fractures, enabling defining at least one, or a chronology of successive fracturing sets to be modelled. In embodiments, such chronology may be a geological calendar of the formation process of the reservoir to be modelled. With reference to figure 3, is shown an example of an excerpt of a geological calendar of the formation process of a reservoir that may be used within the scope of the present disclosure.

A geological calendar is a hypothetical timeline that extends far beyond human history, and is used to represent the major events in Earth’s evolution. A geological calendar typically divides Earth’s history into several eons, which are further subdivided into eras, periods, epochs and ages. The most recent eon, the Phanerozoic, is the one in which complex life forms, including humans, have evolved. It is divided into the Paleozoic, Mesozoic and Cenozoic eras which is turn, are further divided into periods. For instance, the Mesozoic era is divided into the Cretaceous, Jurassic and Triassic periods, which are further divided into epochs and ages. The exemplary excerpt of figure 3 shows the division in ages of some epochs of Paleogene and Cretaceous eras. These various divisions of Earth’s history, namely eons, eras, periods, epochs and ages, are further associated to respective geological times defining the beginning and the end of each division.

Within the context of the present disclosure, the geological calendar of the formation of the reservoir further comprises a plurality of geological events having occurred during the formation of the reservoir and having contributed to shaping the reservoir according to its present-day state. Each geological event is associated to at least one geological time, at which the event occurred. The geological time is typically expressed in millions of years before present. For events spreading over a long period of time, i.e. typically more than one million years, a geological event may be associated to a geological time of beginning of the event and a determined duration. The geological events occurring during the formation of the reservoir may be correlated to determined ages or epochs of the evolution of Earth, i.e. may start at the beginning of an age and end at the end of said age or another age.

The geological events in particular comprise deformation stages, during which fractures can appear within the model. The deformation stages are denoted Frac 1 and 2 in the example of figure 3. The method for modelling the reservoir may thus comprise implementing a plurality of fracturing sets, where each deformation stage of the geological calendar corresponds to at least one fracturing set. Moreover, the geological events of the geological calendar may include other types of geological events that may also be simulated between two fracturing sets. Notably, the geological events may include sedimentation events, referred to as Sed1 and Sed 2 in figure 3, and diagenetic events, referred to as D1 in figure 3.

Sedimentation events may be modelled using forward stratigraphic modelling, in particular using the methods disclosed in WO2020/229863 or WO2020/229866 filed by the applicant. Diagenetic events can for instance include eogenesis events, i.e. early diagenetic events occurring soon after sedimentation, or dissolution, and may be modelled respectively using the methods disclosed in PCT/IB2023/000437 and PCT/IB2023/000218.

When the method comprises implementing the successive geological events in accordance with the calendar, the same model of the reservoir is used for successively modelling each geological event, with the outputs of an event that are used in the setup of the next event. In particular, when the method comprises modelling in sequence a deformation event and a diagenetic event, the fractures generated within the model during the fracturing sets corresponding to the deformation event are integrated within the model, and used for simulating diagenesis. As fractures generally have increased permeability with respect to the surrounding matrix, the accuracy of the fractures disposition may strongly influence the evolution of the reservoir due to diagenesis.

Alternatively or in addition, the received ground data may include at least one of: a Type of fractures to be modelled, a localization of fractures, a density of fractures, geometric parameters of fractures.

The received ground data may be obtained from observation, interpretation and measurements acquired on the reservoir, by the means listed above.

The type of received ground data may vary according to the type of fracture. In particular, regarding geometric parameters, these may include one or more of : one or more orientations of fractures, dimensions or dimensions distributions, dip azimuth, dip angle. Back to figure 1a and 1 b, the method then comprises implementing at least one fracturing set 300 comprising the generation of fault array or fracture corridors within the received model. In embodiments, the method comprises a plurality of iterations of step 300, where each fracturing set 300 corresponds to a respective type of fractures to be generated, and at least one of the sets comprises generating fault arrays or fracture corridors.

The step of implementing the fracturing sets 300 is also performed in accordance with the received ground data. Depending on the type of received ground data, said data may be taken into account for the definition and chronology of the fracturing sets (in particular when the received ground data comprises chronological information regarding the formation of fractures), or for the definition of parameters of the fractures generated during a fracturing set.

Each fracturing set 300 comprises receiving input parameters 310, where the input parameters may comprise: a type of fractures to be generated, and parameters relative to the fractures to be generated.

The input parameters may be provided by a user.

The received input parameters may depend on the selected type of fractures to be generated. Indeed, some fracture types may be readily associated with one or more geometric parameters, such as, but not limitatively, a dip angle of the fractures. However, other parameters may be freely selected by a user, optionally within a predefined range. In addition, some input parameters may also correspond, or be derived from the received ground data.

Generally, all fractures and fractures segments are modelled as bounded surfaces, extending transversally to at least one bed of the model, and which define two opposite walls separated by the fracture.

The received input parameters for the fractures may include one or more of a localization of the fractures, a density of the fractures, and geometric parameters of the fractures. The geometric parameters may include one or more of the following parameters, which are schematically represented in figure 4, showing a single wall siding a fracture plane: strike direction, where strike refers to the line formed by the intersection between a horizontal plane and the fracture plane, dip angle, where the dip is a vector perpendicular to the strike line and extending along the fracture plane, and dip angle is the angle between the horizontal plane and the fracture plane, measured perpendicular to the strike line down to the fracture plane, dip azimuth, which is the angle formed between the dip and the North, length along the main direction of the fracture plane, height, which may be measured either along the fracture plane, orthogonally to length, as shown in figure 4, or vertically, aperture, which is the distance between the two walls siding the fracture plane aspect ratio, which is the ratio of the length of the fracture to its height or the ratio of the length to its aperture.

For fracture corridors or fault arrays, the geometric parameters may also include an orientation of the fractures or faults with respect to a general direction of the corridor or array. Also, for some types of fault arrays, the received geometric parameters may relate to distinct families of fault segments, respectively.

In embodiments, and as shown in figure 1 b, said step of receiving input parameters 310 may be performed once for all fracturing sets 300 to be modelled. In that case, step 310 corresponds to a step of setup of the sequence of fracturing sets to be modelled, in which all fracturing sets are defined by the type of fractures to be generated, the associated parameters and relevant ground data used for the generation of fractures of the set.

With reference to figure 5, the step 320 of generating fracture arrays or fracture corridors from the received ground data and input data will now be described.

Said step comprises generating a plurality of fault array or fracture corridor seeds 321 according to the received input parameters and/or ground data.

These may include location of fault arrays or fracture corridors to be modelled, a density of fault arrays or fracture corridors to be modelled, and, within each fault array or fracture corridor, a density of faults. In embodiments, the densities may be derived from ground data, by selecting observable zones, such as geological outcrops, counting fractures of each type and inferring therefrom density values. Both densities may be constant or variable in space along at least one direction.

The generation 321 of fault array or fracture corridor seeds then comprises implementing a stochastic process, in particular a Poisson process, to randomly draw fault array or fracture corridor seeds within the model according to the received density.

Then, for at least one fault array or fracture corridor seed, the method comprises iterating the following steps.

First, a volume of the fault array or fracture corridor is generated during step 322, based on the position of the fault array seed, and according to received geometrical parameters. The received geometrical parameters may include a length, height and width of the geometrical volume corresponding to the fault array or fracture corridor, and optionally an orientation of said volume.

Then, a plurality of fault segments seeds are generated 323 within the volume corresponding to the fault array. It is underlined that, within the present disclosure, the terms “fault segment” designate an elementary component of a fracture or network of fractures. Therefore, despite using the word “segment”, a fault segment relates to a surfacic i.e. two-dimensionnal element. The generation of fault segments seeds is performed using a stochastic process, in particular a Poisson process, according to the received density.

The generation process then comprises iteratively generating fault segments 324 from the fault segment seeds.

The generation of fault segments from segment seeds may comprise steps of generating a fault segment 324a from a seed, optionally managing intersection 324b between the generated fault segment and a previously generated fault segment, and optionally generating 324c a stress shadow zone around the generated fault segment.

The stress shadow zone may be an elliptical zone that is centered on the fault segment seed and has a length (i.e. major axis of the ellipse) corresponding to the length of the fault segment. The width of the stress shadow zone may be pre- determined or defined by the user. If stress shadow zones are generated, then, for a next iteration of the step 324, a next fault segment is grown from another seed only if said seed is not within the stress shadow zone of a previously generated fault segment of the same set. If a stress shadow zone is generated around a fault segment, a fault segment of the same set developed after will terminate its growth upon reaching the stress shadow zone of the previously generated fault segment.

In embodiments, and with reference to figure 6, the generation of fault segments also takes into account previously generated fault arrays, in order to manage abutting between a fault array and previous fractures. Accordingly, when a fault segment seed of a current fault array is selected for growing a fault segment, said segment is only developed from the segment seed if a line between said segment seed and the current fault array’s seed does not intersect a fracture (which may be a fault segment of a fault array, of any type of fracture described herein) generated during a previous fracturing set. The upper part of figure 6 thus represents a case of intersection between said line and a previously generated fault array, while the lower part represents a case of intersection between said line and a previously generated single fault.

The generation of fault segments 324, the implementation of the steps of managing intersection between segments of a same fracturing sets, and the association of a stress shadow zone to a fault segment may vary depending on the type of fault array or fracture corridor that is modelled.

Fault arrays

As mentioned above, the method enables modelling various types of fault arrays, including relay faults, anastomosing faults and en-echelon faults.

Each type of fault array may be defined by at least one of the following parameters: a number of fault segment families, a density of fault segments, optionally defined for each fault segment family, geometric parameters associated with each fault segment family, and whether or not a fault segment is associated with a stress shadow zone.

Moreover, when a selected type of fault array to be modelled is associated with at least two fault segment families, it is further associated with an order of generation of the fault segments according to their families, and intersection rules between fault segments. Said intersection rules may comprise specific rules depending on the respective families of the intersecting segments, i.e. for instance rules regarding intersection of segments of the same family, and rules regarding intersection of segment of a first family with segments of other families.

Fault arrays: relay faults

With reference to figure 7a, the generation of fault segments 324 of relay faults within a fault array volume will now be described. Relay faults are modelled using two fault segments families, which are associated to respective geometric parameters. Fault segments of a first family extend in a first direction, and fault segments of the second family extend in a second direction transverse to the first. The first direction, and the angle between the first and second directions, are geometric parameters which may be defined by the user and received at step 310, or derived from ground data received at step 200. In embodiments, the angle between the first and second directions is predetermined, for instance equal to 90°.

The geometric parameters for relay faults may further comprise geometric parameters of the first segment family, including an angle of the first segment relative to the array direction, a dip azimuth, a dip angle, a length, a height, an aperture and an aspect ratio. The angle between the first segments and the array direction may be set by a user.

The second segment family corresponds to connectivity segments, which are formed between two first family segments, to connect said segments.

Thus, the generation of fault segments comprises generating all fault segments of the first family based on the generated fault segment seeds. Said first family fault segments are associated with a stress shadow zone, and hence a segment is only generated if the corresponding seed is not within the stress shadow zone of a previously generated segment. Moreover, a segment terminates if it reaches the stress shadow zone of a previously generated segment.

Then, connectivity segments are generated, between two adjacent first family segments (i.e. when the two segments overlap at least in part), when the distance between the two fault segments is below a determined threshold. Each connectivity segment abuts against the two first family segments between which it extends.

Fault arrays: anastomosing faults

With reference to figure 7b, the generation of fault segments of anastomosing faults within a fault array volume will now be described.

Anastomosing faults are modelled using two fault segment families, where the two fault segment families may share the same parameters of length, height, aperture, dip angle (equal e.g. to 90°) and aspect ratio. However, the two segment families extend in two different directions which are transverse to one another. In embodiments, the two segment families have a symmetric orientation with respect to the main direction of the array, where said main direction may be input by a user or derived from ground data. In other words, all the segments corresponding to the first segment family extend along an angle +a with respect to the main direction of the array, where a is preferably less than 45°, and all the segments corresponding to the second segment family extend along an angle - a with respect to the main direction of the array. The direction of each fault segment may however vary within a tolerance margin around said angle or +/- a.

The generation of fault segments seeds in this case comprises performing a Poisson process for each family of fault segment, each Poisson process being performed according to the received density.

The generation of fault segments then comprises generating alternatively a fault segment of the first family, and a fault segment of the second family. The fault segments are developed from the respective segment seed, which is centered on the fault segment. In case of intersection with another segment of the same fracturing set, a newly generated segment always abuts against the previously generated segment. The fault segments are not associated with any stress shadow zone.

Fault arrays: en-echelon faults

With reference to figure 7c, the generation of fault segments of en-echelon faults within a fault array volume will now be described. En-echelon faults are modelled using a single fault segment family, having a constant orientation that is defined by a fixed angle with respect to the direction of the array. The dip angle may also be set, equal for instance to 90°. The parameters received at step 310 may include the length, height, aperture and aspect ratio of the fault segments.

The generation of fault segments thus comprises generating fault segments 141 which are all parallel to one another, and are aligned according to a direction that is parallel to the main direction of the array. For instance, a first seed may be randomly generated, then an axis is generated from the position of the seed and the direction of the array, and the remaining seeds are constrained to be generated on that axis. As all segments are parallel, there is no need to manage intersection with previously generated faults of a same set, but the fault segments generation comprises managing possible intersection with fractures from previous sets, according to the corresponding rules. Moreover, en-echelon faults are modelled without any stress shadow zone.

Fracture corridors

In addition to what has already been described above regarding generation of fault arrays or fracture corridors, a specificity of the latter is that a minimum spacing may be provided between two fracture corridors. The minimum spacing may be predefined or set by the user. The minimum spacing is ensured at the step of generating a volume for a fracture corridor from a seed, by determining the distance between the considered seed and the seeds from which fracture corridors volumes have already been generated. When the distance is below the determined spacing, the considered seed is deleted and no fracture corridor volume is generated therefore.

When fracture seeds are generated within the corridor volume, fractures are iteratively developed 324a from the seeds. A seed is randomly chosen among the seeds within the corridor volume, and a fracture is grown from said seed based the received geometrical parameters (length, height, width), with an orientation that is drawn randomly within a range centered on the orientation of the corridor. The range may be +/- 10° with respect to the orientation of the corridor. The development of a fracture may comprise managing intersections 324b between the currently developing fracture and a previously generated fracture from the same set or previous set, such that two intersecting fractures of a same set cross each other and the intersection of a fracture with a fracture from a previous set may be managed according to pre-defined intersection rules. Last, fractures within fracture corridors do not have a stress shadow zone.

Back to figures 1 a and 1 b, the model obtained once the one or more fracturing sets 300 have been implemented is enriched, as compared to the initial model, with a plurality of fractures, which possibly intersect with one another.

The method may then comprise assigning 400 a geological attribute value to each fault. The geological attribute value may be a petrophysical parameter, for instance a permeability, transmissibility value or a transmissibility reduction coefficient for the core zones of damage zones.

For instance, the permeability value may be a function of the aperture, and the considered fluid that would flow within the fractures. For instance, a permeability value may be assigned to a fault according to the following equation:

Where b is the aperture of the fracture, p is the dynamic viscosity, p is the density of the fluid, g is the acceleration due to gravity, and v is the kinematic viscosity of the considered fluid.

Moreover, the method may also comprise processing 500 the network of fractures in order to convert it into a suitable format for later applications.

In embodiments, step 500 may include computing a pseudo-equivalent fracture 510 from a plurality of fault segments, to render the network of fractures less computationally intensive. Regarding fracture corridors, as schematically represented in figures 9a and 9b, a single pseudo-equivalent fracture may be computed to represent all fault segments within the fracture corridor.

With reference to figure 1a, computing a pseudo-equivalent fracture from a set of fault segments comprises a substep 511 of computing an equivalent permeability Keq corresponding to the fault segments. This step may e.g. be performed by application of the teaching of [Oda, 1985]. The pseudo-equivalent fracture may be associated with a permeability value equal to said equivalent permeability Keq. Computing a pseudo-equivalent fracture then comprises a substep 512 of designing the pseudo-equivalent fracture, i.e. attributing geometrical parameters to the fracture.

Regarding fracture corridors, a pseudo-equivalent fracture may be centered on a seed corresponding to the corridor seed, and exhibit a length and height that are those of the fracture corridor. The pseudo-fracture’s strike direction corresponds to the corridor’s azimuth, and the pseudo-fracture’s dip is set equal to 90°.

Regarding fault zones, a pseudo-equivalent fracture may be centered on the gravity center of each side of the damage zone. The pseudo-equivalent fracture length, height, dip azimuth and dip angle may be set as those of the core zone, respectively.

Last, the pseudo-equivalent fracture is also assigned an aperture, which may be derived from the permeability and surface of the fracture by the following equation: b³.S

^K ~ 127

Where b is the aperture of the pseudo-equivalent fracture, K its permeability, S its surface, and V the total volume represented by the fracture i.e. the fracture corridor or the part of the damage zone located on one side of the core zone, depending on the fault segments that are replaced by the pseudo-equivalent fracture.

In embodiments, step 500 may also comprise triangulating the surfaces of at least some of the generated fractures. Step 500 may also computing 520 a mesh within the obtained model, the mesh being formed of a plurality of three-dimensional cells, and conforming to the surfaces representing the fractures. The three-dimensional cells may for instance be hexahedrons or tetrahedrons.

Alternatively, the processing 500 may comprise generating a graph 530 representing the network of fractures, the graph comprising a plurality of nodes and connections between the nodes. The nodes may then be formed by the seeds of the fractures, and be associated with all the parameters of the corresponding fractures. The connections between the nodes may be formed between two fractures which intersect each other.

Even though figure 1 a shown steps 520 and 530 as alternatives with respect to step 510, the method may encompass a step 510 followed by either of steps 520 and 530. The obtained representation of the network of fractures may be displayed for an interpreter to visualize and analyze the reservoir therefrom. Alternatively, further processing may also be performed on the model integrating the network of fractures. As mentioned above, the method may then comprise modelling diagenetic events in the obtained model. In particular when the network of fractures are represented as a graph, the teachings of PCTIB2023/000217 for modelling dissolution within said network are applicable.

Optionally, the method may also comprise a step 600 of upscaling the obtained network of fractures. Upscaling is a computation step that aims at deriving information and properties from a smaller or more detailed scale towards a larger, coarser scale. It enables in particular to increase the computational efficiency of the model.

Various upscaling methods known to the skilled person may be performed on the obtained network of fractures. The upscaling may be implemented to output a single equivalent medium, where this medium may either represent only the network of fractures, or may represent a medium that is equivalent to the rock matrix and the network of fractures that are within said matrix. Alternatively, the upscaling may be implemented to output a dual media including an upscaled equivalent rock matrix and an upscaled equivalent network of fractures. Some upscaling softwares known to the skilled person are available for instance in the Gocad-Skua suite, or Petrel marketed by SLB,etc.

Once upscaled, the obtained model of the geological reservoir may be used as known by the skilled perform for performing computations and simulations based on fluid flows within the reservoir.

REFERENCES

[Peacock, 2016] Peacock et al. “Glossary of fault and other fracture networks”, in Journal of Structural Geology, Volume 92, 2016, Pages 12-29, ISSN 0191-8141

[Libby, 2019]: Libby, S., et al, “Grown Discrete Fracture Networks: a new method for generating fractures according to their deformation history”; June 2019, ARMA 53^rd US Rock Mechanics / Geomechanics Symposium.

[Oda, 1985]: Oda, M. et al. "Permeability tensor for discontinuous rock masses”, Geotechnique, Vol. 35, Issue 4, pp483-495, 1985.

Claims

1. A computer-implemented method for modelling a reservoir comprising fault arrays or fracture corridors, comprising: obtaining an initial model of the reservoir (100) and ground data (200), Implementing (300) at least one set of fault arrays or fracture corridors generation in the initial model, comprising receiving input parameters (310), and, based on the received ground data and input parameters: o generating a plurality of array or corridor seeds (321 ) in the initial model, and, o For at least one seed,

■ Generating (322), from the seed, a volume of the fault array or fracture corridor,

■ Generating (323), within said volume, a plurality of fault segments seeds, and

■ Generating (324) fault segments from at least part of the fault segments seeds.

2. The computer-implemented method according to claim 1 , wherein the ground data comprises at least one of: a localization of fractures, a density of fractures corridors, fault arrays, or of fractures within a fracture corridor or a fault array, geometric parameters of fracture corridors, fault arrays or fractures within a fracture corridor or a fault array, a chronology of the formation of fractures.

3. The computer-implemented method according to claim 1 or 2, wherein generating (321 ) the array or corridor seeds is performed by implementation of a probabilistic process based on a predetermined density.

4. The computer-implemented method according to any of the preceding claims, wherein generating (323) fault segments seeds is performed by implementation of a stochastic process based on a predetermined density.

5. The computer-implemented method according to any of the preceding claims, comprising implementing at least two sets (300) of fault arrays generation, and the method comprises developing fault segments from a segment seed when a line joining the fault segment seed and the seed of the array to which belongs the fault segment-does not cross a fault segment of a fault array of a previous set.

6. The computer-implemented method according to any of the preceding claims, wherein the generated fault arrays are selected among a plurality of types of fault arrays, where each type of fault arrays is defined by:

A number of fault segments families,

7. The computer-implemented method according to any of the preceding claims, comprising at least one set (300) of generation of relay faults, wherein relay faults comprise two fault segment families associated to respective geometric parameters, fault segments of a first family extending in a first direction, and fault segments of the second family extending in a second direction, transverse to the first direction, and developing fault segments (324) from the fault segments seeds comprises:

8. The computer-implemented method according to any of the preceding claims, comprising at least one set (300) of generation of anastomosing faults, wherein anastomosing faults comprise two fault segment families associated to respective orientation parameters, fault segments of a first family extending in a first direction, and fault segments of the second family extending in a second direction, transverse to the first direction, and developing fault segments (324) from the fault segments seeds comprise generating alternatively a fault segment of the first family and a first segment of the second family, wherein, in case of intersection between a currently generated fault segment and a previously generated fault segment, the currently generated fault segment abuts against the previously generated fault segment regardless the family of the previous fault segment.

9. The computer-implemented method according to any of the preceding claims, comprising at least one set (300) of generation of en-echelon faults, wherein the fault segments seeds are generated along a line extending parallel to the main direction of the array, and all fault segments generated from the fault segments seeds extend parallel to each other.

10. The computer-implemented method according to any of the preceding claims, further comprising a step of assigning (400) at least one of an aperture or permeability value to the generated fault segments.

11 . The computer-implemented method according to any of the preceding claims, further comprising a step of upscaling (600) the obtained reservoir model or the obtained fault arrays.

12. A computer program product comprising code instructions for implementing the method according to any of the preceding claims, when it is executed by a computer (1 ).

13. A non-transitory computer-readable storage having stored thereon code instructions which, when executed by a computer (1 ), cause said computer to implement the method according to any of the preceding claims.

14. A computer (1 ), configured for implementing the method according to any of claims 1-11.