method for seismic analysis of hydrocarbon systems

专利摘要:
METHOD FOR SEISMIC ANALYSIS OF HYDROCARBON SYSTEMS Method to analyze seismic data representing a subsurface region for the presence of a hydrocarbon system or a play (a conceptual model of a hydrocarbon accumulation style often used to develop prospecting in a basin) particular. Seismic attributes are computed, the attributes being selected in relation to the classic elements of a hydrocarbon system, that is, reservoir, sealing, trap, source, maturation, and migration. Preferably, the attributes are computed across tissue structures (1) of the subsurface region, and are attenuated by at least tens or hundreds of voxels of data. The resulting geological attributes (2) are used to analyze the data for elements of the hydrocarbon system and / or recognition of specific plays, and to classify and annotate divided regions (3) of the data volume based on size, quality, and confidence in predicting prospects (5). A catalog (80) of hydrocarbon trap configurations can be created and used to identify the potential presence of hydrocarbon trap and / or assist in the score (4) and / or classification of the divided regions as hydrocarbon prospects.
公开号:BR112012028653B1
申请号:R112012028653-6
申请日:2011-04-22
公开日:2020-11-10
发明作者:Matthias G. Imhof；Pavel Dimitrov；Kelly Wrobel；Krishnan Kumaran；Martin J. Terrell；Stefan Hussenoeder
申请人:Exxonmobil Upstream Research Company；
IPC主号:

专利说明:

[0001] This application claims the benefit of U.S. Provisional Patent Application 61 / 349,534, filed on May 28, 2010, entitled SYSTEM FOR SEISMIC HYDROCARBON SYSTEM ANALYSIS, the entirety of which is incorporated herein by reference. FIELD OF THE INVENTION
[0002] This invention refers generally to the field of geophysical prospecting and, more particularly, to the interpretation of seismic data. Specifically, the description describes a method for detecting and classifying potential hydrocarbon opportunities using seismic data. BACKGROUND OF THE INVENTION
[0003] An active hydrocarbon system is defined by the presence of a porous reservoir formation that provides storage space for hydrocarbons, a seal that prevents hydrocarbons from escaping from the reservoir, good trapping geometry, and a source formation that contains a high percentage of biogenic material. Under the influence of high temperature and increased pressure, the biogenic material is matured (or cooked) to form hydrocarbons including gas, crude oil, asphalt and tar. Driven by fluctuation and differential pressures, hydrocarbons migrate and a fraction of these hydrocarbons accumulate in traps formed by accidental geometric arrangements of reservoir formations (ie, trapping geometries) and seals. Traps have a finite volume, however, and can spill or leak some of the accumulated hydrocarbons, the part of which can then accumulate in other tanks.
[0004] Seismic images of the subsurface allow interpreters to identify some potential traps based on practical and suggestive geometries. Sometimes, seismic data can provide a direct indication of the presence of hydrocarbons. Standard interpretation practices, however, are labor intensive and often focus on areas where the interpreter collects some indication of prospects. Many opportunities, therefore, remain undetected because the indications are very subtle or hidden, for example, by seismic noise. Even if signs of foresight are observed, they may not be examined when in the presence of the most obvious opportunities, or when the interpreter is limited by time constraints. Thus, some hydrocarbon accumulations are discovered late or remain undiscovered.
[0005] Published experiences to solve similar problems include the following:
[0006] “Method for Seismic Interpretation Using Seismic Texture Attributes” (PCT Patent Application Publication WO 2010/053618), by Imhof, describes a method for computing texture attributes that can be used for scoring and segmenting seismic data based on your local appearance. The texture can be used to define seismic facies.
[0007] “Windowed Statistical Analysis for Anomaly Detection in Geophysical Datasets” (PCT Patent Application Publication WO 2010/056424), by Kumaran et al., Describes a method for examining seismic data in regions that are statistically anomalous in the context of the data and thus, they serve to statistically highlight unusual or salient areas.
[0008] “Method For Geophysical and Geological Interpretation of Seismic Volumes In Depth, Time, and Age” (US Patent Application Publication No. 2010/0149917), by Imhof et al., Describes a method for transforming seismic data from geophysical domains of depth or time of displacement from two directions to a domain of geological age in which all seismic reflections are approximately horizontal and comparable to their state in the geological time of their formation. Seismic attributes formed from this age domain can improve the definition of the elements of the hydrocarbon system.
[0009] European Patent EP1110103 B1 ("Method Of Seismic Signal Processing"), Meldahl et al., Describes a method for finding areas in seismic data that are similar to the areas specified by the interpreter. In addition, she describes the use of this method to find seismic indications of percolating hydrocarbons.
[0010] US Patent No. 6,226,596 B1 ("Method for analyzing and classifying three dimensional seismic information"), by Gao, describes a method for generating seismic texture attributes that can be used for segmentation, score, or definition of seismic facies.
[0011] US Patent No. 6,438,493 B1 (“Method for seismic facies interpretation using textural analysis and neural networks”) from West and May, describes a method for generating seismic texture attributes that are used in a supervised score to designate seismic facies attributes.
[0012] US Patent No. 6,516,274 B2 (“Method for imaging discontinuities in seismic data using dip-steering”) by Cheng et al., Describes a method for detecting edges or discontinuities in seismic data that often indicate rupture and displacement of faults in the subsurface of the regular layer.
[0013] US Patent No. 6,725,174 B2 (“Edge-preserving enhancement of seismic images by nonlinear anisotropic diffusion”) to Bouts et al., Describes a method of processing seismic data that accentuates seismic data by removing noise disconnected, while preserving pronounced discontinuities corresponding to flaws or rapid stratigraphic transitions, such as entrenched channels.
[0014] US Patent No. 6,850,845 B2 ("System for multi-dimensional data analysis"), by Stark, describes a method of processing seismic data that allows the computation of a geological time attribute used for derivative attributes of planing and associated.
[0015] US Patent No. 6,850,864 B2 ("Method for analyzing dip in seismic data volumes"), by Gillard et al., Describes a method of processing seismic data for estimating reflection slopes that define the structure local.
[0016] U.S. Patent No. 7,203,342 B2 ("Image feature extraction"), to Pedersen, describes a method of processing seismic data to detect and improve gaps or horizons in relatively noisy data.
[0017] US Patent No. 7,424,367 B2 ("Method for predicting lithology and porosity from seismic reflection data"), by Saltzer et al., Describes a method for predicting lithology and porosity of subsurface rocks from seismic data and, thus, allows the differentiation between reservoir rocks and sealing rock formations.
[0018] US Patent No. 7,454,292 B2 (“Inverse-vector method for smoothing dips and azimuths”), by Wang et al., Describes a method for the robust computation of horizon slopes and orientations that define the local structure .
[0019] Neelamani and Converse PCT Patent Application Publication WO 2009/01 1735 (“Geologic Features From Curvelet Based Seismic Attributes”) describes a method for computing hydrocarbon indicators or texture attributes that can be used for the identification of subsurface aspects.
[0020] PCT Patent Application Publication WO 2009/082545 Al (“Detection Of Features In Seismic Images”), by Kumaran and Wang, describes a method for detecting faults, channels, and similar aspects in seismic data.
[0021] PCT Patent Application Publication WO 2009/137150 Al (“Method For Geophysical And Stratigraphic Interpretation Using Waveform Anomalies”) by Imhof describes a method of processing seismic data to map stratigraphic terminations and pinch outs (thinning of the reservoir layer).
[0022] PCT Patent Application Publication WO 2009/137228 A2 (“Transport Property Data Calculated From Derivative Seismic Rock Property Data For Transport Modeling”), by Oppert et al., Describes a method for evaluating properties, such as flow heat or fluid permeability, which affects elements of the hydrocarbon system.
[0023] PCT Patent Application Publication W02009 / 142872 A (“Seismic Horizon Skeletonization”), by Imhof et al., Describes an automatic method for extracting a large number of horizons from a set of seismic data. In addition, it describes a wide-ranging reconnaissance workflow that splits a data set, analyzes regions, and classifies them according to their potential to contain hydrocarbons.
[0024] “A New Class of Large-scale Attributes for Seismic Stratigraphy”, by Gesbert et al, 71st EAGE Conference & Exhibition, (2009) describes a set of stratigraphic attributes computed from two-dimensional seismic data that accentuate regional non-conformities and trends regional tuning and quantifies trends in regional seismic facies.
[0025] “Applications of plane-wave destruction filters” by Fomel, Geophysics 67, 1946-1960, (2002), describes a method to assess the inclination and orientation of the seismic horizon that define the local structure.
[0026] Imhof's “Estimating Seismic Heterogeneity with the Structure Tensor”, 67th EAGE Conference & Exhibition, (2005), describes a method for assessing the seismic horizon slope and orientation that define the local structure and seismic texture attributes that characterize the local heterogeneity.
[0027] “Flattening without picking” by Lomask et al, Geophysics 71, P13-P20 (2006), describes a method of processing seismic data to approximately level the data that allows item characterization of some elements of the hydrocarbon system.
[0028] “Hydrocarbon leakage interpreted on seismic data” by Loseth et al, Marine and Petroleum Geology 26, 1304-1319, (2009), describes drive-interpreter methods for detecting hydrocarbon filtration through the subsurface.
[0029] “Hydrocarbon Traps, K.T. Biddle and C.C. Wielchowsky, The Petroleum System - From Source to Trap, AAPG Memoir 60, pgs. 219-235, (1994), presents a collection of types of hydrocarbon traps.
[0030] “Imaging Vector Fields Using Line Integral Convolution” by Cabral and Leedom, Proceedings of ACM SigGraph 93, 263-270, (1993), describes a method of visualizing vector fields of flow lines.
[0031] “Lithofacies Prediction in Deep Water Water Reservoirs” by Oppert et al, Society of Exploration Geophysicists, Expanded Abstracts, 1708-171 1, (2006), describes a method for assessing subsurface lithology using seismic and wired data .
[0032] Minutes of the 2008 conference on “Seismic Rock-Property Inversion and Lithofacies Prediction at Erha Field, Nigeria” by Xu et al, Nigerian Association of Petroleum Explorationists (NAPE), describes a method for assessing subsurface lithology using seismic data and wired.
[0033] Randen and Sonneland (“Atlas of 3D Seismic Attributes”, in Mathematical Methods and Modeling in Hydrocarbon Exploration and Production, Iske and Randen (editors), Springer, pages 23-46 (2005)), presents a summary of attributes three-dimensional seismic that characterize aspects of seismic or seismic-stratigraphic texture.
[0034] What is needed is an automated system that explores an entire data set for the elements of a hydrocarbon system and issues a list of prospects for the interpreter to examine. Preferably, this list of potential targets is classified by expected volume, presence and quality of elements of the hydrocarbon system, and confidence in their detection and identification. Preferably, the list of prospects is also noted. The present invention meets at least these needs. SUMMARY OF THE INVENTION
[0035] The invention, in one of its aspects, is a method that computes seismic attributes of multiscale and typically oriented structure that refer to the classic elements of a hydrocarbon system, that is, reservoir, sealing, trape, source, maturation , and migration. The attributes are spatially correlated and compared with a catalog of hydrocarbon trap configurations to determine the potential presence of hydrocarbon trap and to assess the confidence of its existence.
[0036] In one embodiment, the invention is a computer-implemented method for analyzing a volume composed of voxels of seismic data representing a subsurface region as to the presence of a hydrocarbon system or a particular play, comprising: - dividing the volume of seismic data to form a plurality of segments; e - classify the plurality of segments as to the presence of a particular hydrocarbon or play system, based at least partially on the prospective scores for the seismic data voxels in each segment; where the prospective score is based on the computation of at least two attributes that refer to different elements of a particular hydrocarbon or play system.
[0037] As with any geophysical data processing method, the invention in practical applications is highly automated, that is, it is carried out with the help of a computer programmed according to the descriptions here. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] The present invention and its advantages will be better understood by reference to the following detailed description and the accompanying drawings, in which:
[0039] Fig. 1 illustrates elements of the Hydrocarbon System for an anticline trape;
[0040] Fig. 2 is a flow chart showing basic steps in an embodiment of the invention;
[0041] Fig. 3 is a schematic diagram illustrating an embodiment of the present inventive method;
[0042] Fig. 4 illustrates how the filter stencil of an integral line convolution filter follows tangent directions;
[0043] Fig. 5 illustrates how the full line convolution filter can be made more robust;
[0044] Fig. 6 illustrates that the integral thread convolution filter stencil can have an extended thickness.
[0045] Fig. 7 shows an example of converting a local attribute to a regional one (left: seismic data, half-left: local convergence attribute, half-right: tangent vector field, and right: regional convergence) ;
[0046] Fig. 8 shows an example of three-dimensional convergence attributes with a convergence magnitude component (light: strong convergence, dark: no convergence) and a convergence orientation component;
[0047] Fig. 9 defines stratigraphic termination relationships;
[0048] FIG. 10 represents a structural bounded fold (anticline);
[0049] The pig. 11 represents a structural fault related fault;
[0050] FIG. 12 shows a salt-related structural trap;
[0051] pig-15 illustrates a stratigraphic trape related to a pinchout of the reservoir;
[0052] Fig-14 shows a stratigraphic trape related to a non-conformity;
[0053] Fig-15 represents a stratigraphic trape formed by buried erosion relief;
[0054] Fig-16 represents a stratigraphic trape formed by diagenetic differences;
[0055] Fig-17 presents a schematic depositional sequence model, that is, an ingot diagram;
[0056] Fig-18 illustrates how the different elements can be integrated to assess the chance of success for an accumulation of hydrocarbons;
[0057] Fig. 19 shows a schematic application of classification of four potential targets; and
[0058] Fig. 20 illustrates an exemplary application of the present inventive method.
[0059] The invention will be described in combination with the exemplary embodiments. However, insofar as the following detailed description is specific to a particular embodiment or particular use of the invention, it is intended to be illustrative only and is not to be construed as limiting the scope of the invention. On the contrary, it is intended to cover all alternatives, modifications and equivalents that may be included within the scope of the invention, as defined by the appended claims. DETAILED DESCRIPTION OF EXEMPLARY CARE
[0060] The invention, in one of its aspects, comprises computation of seismic attributes that relate to elements of the hydrocarbon system, analysis of data for elements of the hydrocarbon system and / or recognition of specific plays or conceptual styles, and classification and annotation of these regions based on the size, quality, and confidence of the prospects. Some definitions are given below.
[0061] Although the term can be used more widely or strictly elsewhere, an oil or hydrocarbon system is generally used here to mean a natural system that covers an active matrix rock capsule and all related oil and gas. This includes all the geological elements and processes that are essential for a hydrocarbon accumulation to exist, as illustrated in Fig. 1. The hydrocarbons found in reality include high concentrations of thermal and / or biogenic gas found in conventional reservoirs or in gas hydrates, reservoirs tight, fractured shale, or coal; and condensates, crude oils, heavy oils, asphalt and tar. The term "system" describes the independent elements and processes that form the functional unit that creates hydrocarbon accumulations. The essential elements include an oil matrix rock (source), reservoir rock (reservoir), sealing rock (seal), and cover rock (overburden). The processes are formation of the trap and maturation (generation), migration, and accumulation of hydrocarbons. A sequence or distribution of events is implicit in these processes.
[0062] An alternative definition of the hydrocarbon system may include only the matrix rock, the maturation and migration processes, and their distribution; in this case, reservoir, seal and trap can be defined to form a play. For the purpose of explaining the present inventive method, the term hydrocarbon system is defined to encompass source, reservoir, sealing, trap, maturation, migration and distribution. In addition, the term play is generally used here to indicate a specific combination and arrangement of reservoir, sealing and entrapment geometry.
[0063] The matrix rock is a rock rich in organic matter that, if heated sufficiently, will generate oil and / or gas over time. Common matrix rocks include shales or limestones. Rocks of marine origin tend to be oil-prone, while land-based rocks (such as coal) tend to be gas-prone. The preservation of organic matter without degradation is critical to create a rock with a good source, and necessary for a complete oil system.
[0064] The reservoir is a subsurface rock body having sufficient porosity and permeability to receive, store, and transmit fluids. Sedimentary rocks are the majority of common reservoir rocks, because they have more porosity than most igneous or metamorphic rocks and form conditions of lower temperature in which hydrocarbons can be preserved. A reservoir is a critical component of a complete oil system.
[0065] Sealing rock is a relatively impermeable rock, usually shale, anhydrite, or salt, which forms a barrier or cover above and partially around the reservoir rock, so that fluids cannot migrate beyond the reservoir. A sealing rock is a critical component of a complete oil system.
[0066] Cover rock is the rock on top of the fountain and reservoir. In the context of the oil system, its main function is to form a thick blanket over the source, where it increases the temperature and pressure to the degree necessary to convert organic matter into hydrocarbons.
[0067] Trape is a configuration of rocks suitable for containing hydrocarbons and is sealed by a relatively impermeable formation through which hydrocarbons will not migrate. Traps are described as structural traps (in deformed strata, such as folds and faults) or stratigraphic traps (in areas where rock types change, such as non-conformities, pinch-outs and reefs) or their combinations. For structural traps, deformation must occur before hydrocarbon migration, or the hydrocarbons will not accumulate. A trap is an essential component of an oil system.
[0068] Generation or maturation is the formation of hydrocarbons from a matrix rock, since bitumen forms kerogen and accumulates as oil or gas. The generation depends on three main factors: presence of sufficient organic matter to produce hydrocarbons, adequate temperature, and sufficient time to bring the matrix rock to maturity. The pressure and presence of bacteria and catalysts also affect generation. Insufficient pressure and temperature, caused, for example, by a superficial burial with a thin covering, will yield an immature source and generation will be scarce or incomplete. Excessive pressure and temperature, caused, for example, by deep burial under a thick covering, will cause degradation of the oil generated in gas and, subsequently, in carbon dioxide and water. Generation is a critical phase in the development of an oil system.
[0069] Migration is the movement of hydrocarbons from their source into the reservoir rocks. The movement of newly generated hydrocarbons out of its parent rock is the primary migration, also called expulsion. The other movement of hydrocarbons into the reservoir rock in a hydrocarbon trap or other accumulation area is secondary migration. Migration typically occurs from a structurally low area to a higher area, due to the relative fluctuation of hydrocarbons compared to the surrounding rock. Migration can be local or can occur over distances of hundreds of kilometers in large sedimentary basins and is critical to the formation of a viable oil system.
[0070] Accumulation refers to both an occurrence of trapped hydrocarbons, that is, a play or an oil or gas field, as well as the development phase of an oil system, during which hydrocarbons migrate inward and remain trapped in the reservoir rocks.
[0071] Distribution refers to the relative order in which elements are formed or modified, or the order in which the processes occur. A trap can accumulate hydrocarbons by migrating only if formed before migration. A trap may not be completed if the migration has not yet reached its location. A trap can lose its load, at least partially, if the sealing rock is broken after accumulation.
[0072] A play is a conceptual model for a hydrocarbon accumulation style, often used to develop prospects in a basin, region or trend, or used to continue to explore an identified trend. A play (or a group of interrelated plays) usually occurs in a single hydrocarbon system and can be understood by a group of similar prospects.
[0073] A prospect is an area where hydrocarbons have been predicted to exist in economical quantities. A prospect is often an anomaly, such as a geological structure or an anomaly of seismic amplitude that is recommended as a location for drilling a well to verify economical quantities of hydrocarbons. The justification for drilling a prospect is made by joining evidence for an active hydrocarbon system, or by demonstrating reasonable probabilities of finding good quality reservoir rock, a sufficiently sized trap, adequate sealing rock, and appropriate conditions for generation and migration of hydrocarbons to fill the reservoir. For the purposes of the inventive method, the prospectus is used widely to indicate an area that is recommended for further detailed analysis.
[0074] As mentioned above, what is needed is a method that automatically analyzes seismic data for the presence of elements of the hydrocarbon system, flag regions where play elements are juxtaposed in favorable configurations or consistent with a known or specified play, and classify these prospects with reference to their hydrocarbon accumulation potential. Such a system focuses on analysis and interpretation in more prospective areas. In addition, the system can recognize the type of play and provide a confidence score for individual elements. For each prospect, unidentified elements or elements with low confidence are vital and require special attention during subsequent analyzes to avoid exposing the prospect to danger. The system can be used multiple times during the life cycle of a region or asset. First, the system can be used in regional data, typically two-dimensional, to identify prospecting areas, for example, to prepare an offer for a block or to locate a three-dimensional exploration survey. In the exploration phase, the system can be used to examine carefully, to focus and guide the interpreter to smaller and more controllable subsets of the volume of seismic data. At the production stage, the system can be used to locate smaller prospects close to previously located infrastructure. Finally, the system can be used to ensure that no prospectus has been inspected, before abandoning or trading the asset.
[0075] Fig. 2 is a flow chart showing basic steps, in an embodiment of the inventive method, which include two determined steps (boxes having continuous lines) and six optional steps (boxes with dashed lines) that depend on inputs, assumptions made , and the way in which the system is employed. In step 2, attributes are generated that refer to elements of the hydrocarbon system or specific play elements. In step 5, the data is analyzed and locations are classified with respect to the attributes, thus identifying potential prospects that are then stored for another analysis or visualization.
[0076] Optional Step 1 defines a subsurface fabric, that is, the geometries of the subsurface layers and their deformation by failures and folds, which form the basis for computing some attributes or may be necessary for other computations, to aggregate information from similar strata. Many attributes, which are useful for the inventive system, are formed by integrating or averaging more traditional attributes across the fabric. In optional Step 3, the data is separated into at least two segments or divisions for analysis and prospecting. Optional step 4 is the formation of normalized scores for elements of the hydrocarbon system using one or more of the attributes. Optional step 6 is the analysis of the formed and classified prospects; for example, identification of the smallest element (s) expressed that has yet to be examined. Step 7 is the definition or selection of at least one type of specific concept or play that is to be searched for instead of a generic search for neighboring elements of the hydrocarbon system. Finally, step 8 is the definition, creation and control of a catalog of types of play and the configuration of its elements.
[0077] A schematic application of the inventive method is presented in Fig. 3, in which seismic attributes 32 are computed from seismic data 31. The attributes are then combined 33 in highlighted regions (for example, 36) that can constitute stratigraphic plays , in this case the cover-up play 34 and the stratigraphic play pinch-out 35.
[0078] Definition of Fabric
[0079] An attribute is a measurable property of seismic data, such as amplitude, slope, frequency, phase and polarity. Attributes can be measured in an instant of time or through a time window, and can be measured on a single trace or set of traces, or on an interpreted seismic data surface, a data window, or even in multiple volumes. seismic simultaneously. The present inventive method employs many classic attributes, which are well known to those skilled in the art of analyzing attribute or seismic interpretation. For the inventive method, however, some of these attributes are modified, for example, integrating along strata or structure, to emphasize regional variations across certain locations. In addition, new geological attributes (2) are described that directly refer to elements of the hydrocarbon system or play elements, and to particular configurations of these elements. Since geological attributes can be covered in alternative modes, and different geological attributes can refer to the same element or configuration of elements, scores (4) can be computed that synthesize the different forms of realization and attributes among themselves for a measure that assess the chance of finding a particular play element or hydrocarbon system, or a particular geometric arrangement of such elements. The scores are then combined (6) into probabilities for the settings that can be used for viewing, analysis, or rating and rating.
[0080] With respect to Fig. 2, less than all the steps illustrated may be required to implement a particular embodiment of the invention. The individual steps can be combined or separated into multiple components. In addition, additional and / or alternative methodologies may employ additional steps not shown here. Although the flowchart illustrates several actions taking place consecutively, it should also be noted that some actions could take place in series, substantially in parallel, and / or at considerably different time points. The steps can also be repeated. An example of these variations is step 1, the definition of fabric, structure or layer structure.
[0081] The fabric may be necessary for some attributes used in the invention and, thus, when using only such attributes, the definition of fabric can be ignored. Most of the attributes used for the inventive method, however, will require a fabric for their formation. The fabric could be computed in motion for the entire attribute when needed. Preferably, however, the fabric is generated only once, then stored and used repeatedly to generate, modify, or integrate attributes. In Step 1, tissue formation is shown as an optional step, because, depending on the attributes used in the different embodiments of the inventive method, the tissue should be computed once, repeatedly, or in no way.
[0082] One advantage of generating tissue once is the consistency between different attributes, because they are all based on the same tissue. A disadvantage of generating it only once is that the resulting tissue needs to be represented and stored in some way that may be inefficient or even inappropriate for some applications of generation, modification, or attribute integration. Alternatively, each application can generate its own tissue in an adequate and effective representation that can cause deficiencies in the total system, because the same tissue with the same representations can be computed multiple times. Another option is to compute the fabric with some chosen methods and representations, which are stored and thus reusable. With this alternative method, the disadvantages are increased storage requirements and repeated access to the storage media. In practice, the preferred method may be to compute and store the tissue or its components for some of the most common methods and representations, and to compute less tissue when necessary.
[0083] The methods for computing the tissue can be based on estimates of inclination, orientation or collision; structure tensors, waveform correlations, or skeletonizations. A particular way of estimating slope and slope orientation is by using gradients in seismic data in both horizontal and vertical directions (for example, US Patent No. 6,850,864 B2, “Method for analyzing dip in seismic data volumes” Gillard et al). Another slope estimator is based on flat-wave deconstruction filters (for example, “Applications of plane-wave destruction filters” by Fomel, Geophysics 67 (6), 1946- 1960, (2002)). Methods based on the structure tensor allow the computation of normal and tangent vectors in seismic reflections (for example, “Estimating Seismic Heterogeneity with the Structure Tensor” by Imhof, 67th EAGE Conference & Exhibition, (2005); or US Patent No. 7,454., 292 B2, “Inverse-vector method for smoothing dips and azimuths”, by Wang et al). Slopes can be resolved in time shifts for automatic surface picks (“Flattening without picking”, by Lomasket al, Geophysics 71 (4), P13-P20 (2006)).
[0084] An alternative method of computing slope and reflection azimuth is based on the isocontour gradient. The gradient is a local vector that characterizes the steepest incremental direction of a function. Considering seismic data, amplitude, for example, as a function f in three-dimensional space (x, y, t), the gradient is defined as Vf. Since there are three components in the resulting gradient vector Vf, only three neighboring samples are needed to evaluate Vf for the first order. Using more surrounding samples, however, it is possible to estimate the gradient of an overdetermined system, for example, in the sense of the minimum squares, or by using a higher order approximation for the finite difference approximation. In either case, the resulting gradient will be more uniform. The selection of neighboring points is not limited to any particular sequence or pattern. Preferably, the neighborhood can be defined as points or voxels arranged in a sphere, a cylinder, a box, or any other type of shape surrounding the point of interest. In addition, this shape can be aligned with the fabric or with an estimate of the entire point. Derivative attributes can then be derived from this gradient, including magnitude and direction, that is, slope and azimuth in 3D. The gradient, as ordinarily defined, always points in the direction of increasing amplitude. In seismic data with amplitude cycling from negative valleys to positive peaks and vice versa, the result would be that the parallel layers would exhibit opposite slope values. One solution is to normalize the gradient with respect to a mid-space by turning in your direction, for example, to reverse the gradient sign, so that all gradient points are oriented in the positive y direction:

[0085] This gradient calculation is a more robust formulation than the existing gradient calculation methods. It also has the ability to multi-scale, since the window size (number of neighboring points), within which the gradient is computed, can be varied.
[0086] The lateral correlation of waveforms is another method of generating tissue. In a first step, events, such as peaks, valleys, and / or zero crossings, are determined. In a second step, waveforms around these events are compared with the events of neighboring features to find and connect the most similar ones. We observe the asymmetry that occurs due to the fact that each event connects only with another event, that is, the most similar to it. Multiple events, however, can connect to one and the same event. The result is a directed graph structure in which the vertices correspond to events, such as peaks, valleys, and / or zero crossings; and borders connect individual events to the most similar of neighboring features. Starting from an event on a particular line is often possible to follow boundaries between events and lines and then return to a different event on the original line, which indicates stratigraphic (or topological) inconsistencies. To distinguish this method from others, the terms gross skeleton or skeletonization will be used, if necessary.
[0087] A preferred method for computing a tissue is by topological skeletonization, which automatically creates reflection-based surfaces in a topologically consistent manner, in which individual surfaces do not overlap and sets of multiple surfaces are consistent with stratigraphic overlapping principles (Publication of PCT Patent Application WO 2009/142872 Al, “Seismic Horizon Skeletonization”, by Imhof et al.). To distinguish this method from others, the terms topological skeleton or topological skeletonization are used, if necessary. In this method, topologically consistent surfaces are monotonically labeled in a top-downward model, which allows the designation of a pseudoity for all samples of seismic data and the conversion of seismic data from the traditional geophysical depth or time-of-two domain -directions for a geological age domain (for example, US Patent No. 6,850,845 B2, “System for multidimensional data analysis”, by Stark; PCT Patent Application Publication WO 2009/142872 Al, “Seismic Horizon Skeletonization”, by Imhof et al .; and continues in US Patent Application Publication No. 2010/0149917 "Method for Geophysical and Geological Interpretation of Seismic Volumes in Depth, Time, and Age" by Imhof et al.). A resulting age mapping volume, and / or depth mapping volume, allows the transfer of data between the geophysical depth (or time-of-two-directions) domain and the built geological age domain and vice versa. Each horizontal slice through a volume of depth mapping, corresponds to a depth structure map (or time) for a particular horizon.
[0088] Another skeletonization method is based on morphological thinning; the result will be called a morphological skeleton. Seismic data is blocked for binary images, for example, reducing the data to just its polarities, for example, ± 1. The +1 value ranges are reduced to lines with a +1 value with equal connectivity, for example, by applying morphological thinning. -1 ranges are reduced to lines with a -1 value. All other samples are determined to be zero. A similar result is obtained from the evident polarity attribute that is formed by the instantaneous phase polarity computed at the extreme local amplitude.
[0089] Another aspect of tissue computation refers to flaws that can be detected as discontinuities in seismic data (for example, US Patent No. 6,516,274 B2, “Method for imaging discontinuities in seismic data using dip-steering”, by Cheng et al .; or PCT Patent Application Publication WO 2009/082545 Al (“Detection Of Features In Seismic Images” by Kumaran and Wang). Detected fault segments can still be completely cleaned or refined (for example, ( US Patent No. 7,203,342 B2, “Image feature extraction”, by Pedersen).
[0090] Geological Attributes
[0091] 1st Generation of Regional Attributes of Long Distance Structure Smoothly Oriented
[0092] A geological attribute is a seismic attribute that enhances or quantifies some aspect of the hydrocarbon system or a play. In reality, geological attributes are often regional. They can, for example, estimate a seismic data property using tens or hundreds of voxels. Through such distances, the attributes are preferably computed along the fabric in a structure-oriented manner in order to avoid mixing and, thus, contamination of other strata cuts through the analysis window.
[0093] Given the tissue as defined by the original seismic data volume, any attribute can be converted into a regional attribute, integrating or averaging it across the tissue. One such method is based on non-linear anisotropic diffusion filtering (for example, US Patent No. 6,725,174 B2, “Edge-preserving enhancement of seismic images by nonlinear anisotropic diffusion”, by Bouts et al.), Which flattens the data seismically predominantly along the direction of the reflections. Traditionally, the tissue or the directions of the reflections are computed using seismic structure tensors for the given set of seismic data that is to be flattened, however the tissue can be defined with any method. Some are described here.
[0094] In the present invention, non-linear anisotropic diffusion can be used to compute the tissue for the primary seismic amplitude volume, but then the tissue is used to flatten a set of secondary data along the direction of the reflections from the volume of primary seismic amplitude. In some cases, this secondary data set cannot even be extended in layers or bands, which avoids computing its own structure tensor. In cases where the secondary data set is streaked, it can, however, be advantageous to plan along the fabric of a reference seismic volume, for example, to increase consistency between different data sets after structure planing -oriented.
[0095] An alternative method of oriented structure, planing over long distances of the present invention, is based on integral line convolution (LIC). Integral line convolution is a well-known texture synthesis algorithm used in image processing or data visualization (“Imaging Vector Fields Using Line Integral Convolution” by Cabral and Leedom, Proceedings of ACM SigGraph 93, 263-270, (1993 )). A low-pass filter is used to wrap an incoming interference texture, for example, a random interference image, along symmetrically bidirectional pixel-centered aerodynamic shapes to explore spatial correlation in the direction of flow. Often, aerodynamic shapes are not represented by curves, but by a vector field whose vectors are tangent to aerodynamic shapes, and aerodynamic shapes are computed by integration.
[0096] The replacement of some seismic attribute data for the interference texture, and using the fabric to define the aerodynamic shapes, allows to use integral line convolution to filter the attribute given along the fabric, which is a new approach to filtering structure-oriented. Preferred tissue definitions are based on slope, for example, when computed from structure tensioners. More efficiency is gained by the discretization of the fabric or, more specifically, by the slopes or tangents of reflection, at 0, 45, 90, 135, 180, 225, 270, or 315 °; that is, purely lateral, purely vertical, or purely diagonal, which replaces the numerical integration and interpolation associated with movement along an associated list. Since seismic data is typically more or less horizontally streaked, more efficiency is gained by forcing all tangents to go from left to right, which reduces the slopes to 0, +45, and -45 °. Figure 4 illustrates the application of a line convolution filter in which the gray crosses indicate the location of the samples and the arrows indicate the discretized slope directions. Filters are applied to the locations indicated by the thick black dots. The drawings show how the filter stencils centered on the points follow the vectors and, thus, illustrate the fabric by collecting the sample values, marked with black crosses, along these paths.
[0097] To make the full line convolution filter more robust in the presence of flaws and other complexities, edge preservation can be done by applying the filter in multiple stencil configurations, which are not necessarily centered on the current sample location (illustrated in Fig. 5). Out of all tested stencil configurations, the one with the least variation between your samples is preferable and finally applied in a particular location. Depending on the filter coefficients and the nature of the attribute being treated with such an integral edge preservation line convolution filter, criteria other than the minimum variation may apply. Other possible criteria include maximum variation, minimum or maximum output, and / or avoidance of specially flagged sample locations.
[0098] The filter coefficients need not be uniform, which allows the application of differentiators (high-pass filter), integrators (low-pass filter), or their combinations. Filters can be linear or non-linear, for example, medium filters or mode filters. Figure 6 illustrates that the filter stencil may have an arbitrary shape that is deformed to conform to the vector path through the current sample location. Figure 7 shows an exemplary application of such filtering tools, converting a local attribute into a regional attribute.
[0099] Another method used to perform structure-oriented filtration over long distances involves the exploration of surfaces obtained by seismic skeletonization (PCT Patent Publication No. WO 2009/142872 Al, “Seismic Horizon Skeletonization”, by Imhof et al. ), in which the skeleton surfaces provide the tissue along which the attributes are filtered. Instead of planing along different surfaces, which may require interpolation, planing can be performed along isosurfaces of the age-mapped volume of the skeleton surfaces. Alternatively, the attribute can be transformed by the traditional domain of depth or travel time from two geophysical directions to the domain of geological age. In the domain of geological age, a horizontal slice corresponds to a horizon and, therefore, flattening or filtering of horizontal slices is, by definition, structure-oriented (US Patent Application No. 12 / 623.034 “Method For Geophysical and Geological Interpretation of Seismic Volumes In Depth, Time, and Age ”, by Imhof et al.). After filtering on the age domain, the results can be transformed back to the depth domain or two-way travel time.
[00100] Regardless of the method used for structure-oriented filtering, geological attributes can be computed as multiscale attributes due to the filter size being specified by the interpreter or system. Instead of computing a geological attribute on only one scale, an interpreter or system can choose multiple filter sizes, allowing computation of the same geological attribute on different scales.
[00101] 1.b Hessian Attributes
[00102] A new seismic attribute of the present invention is the Hessian or second-order spatial derivative (dyadic) H = VVf, which is computed in a model similar to that of the modified gradient. In any place of interest, Hessian can be represented in three dimensions by a symmetric 3x3 matrix, that is, by six independent components. To calculate Hessian at any voxel location, at least six neighboring voxels are required, since it has six independent components. The calculation, however, may involve more voxels to obtain flattening and more stable estimates for Hessiaπo: calculating its components from a super-determined system, for example, in the sense of minimum squares, or using a higher order approximation for the finite difference approximation of spatial derivatives. The selection of neighboring points is not limited to any particular sequence or pattern. Preferably, the neighborhood can be defined as points or voxels arranged in a sphere, a cylinder, a box, or any other type of shape surrounding the point of interest. In addition, this shape can be aligned with the fabric or with an estimate of the entire point.
[00103] From a geometrical point of view, the volume of seismic data consists of point aspects (for example, noise), curvilinear aspects (for example, terminations or channels), surface aspects (for example, faults), or aspects of volume (for example, channel straps or saline bodies) embedded in a red background structure. In recent practices, derivatives of second-order directional amplitude or trace correlation signals are generally calculated to detect heterogeneities that typically occur in geological aspects, such as faults, erosion limits, and pinch-outs. Lacking cross-derivatives, attributes based on traditional second-order derivatives incompletely represent local variations of the data. With its six components, Hessian characterizes the local convexity or concavity of the seismic data. Traditional second-order derivatives are found diagonally across the Hessian, while cross-derivatives are found outside the diagonal. The attributes can thus be formed from these six components and their combinations. The components, and therefore the attributes, are formed from second order derivatives of a seismic volume and show spatial change in gradient. They reflect spatial changes in angles of inclination and azimuth of the seismic volume, which allows to highlight anomalous regions and groups of interrelated aspects. By highlighting different types of “discontinuities” and their spatial interrelationships, seismic volumes can be better characterized and visualized.
[00104] Another use of Hessian is to serve as an intermediate step to calculate spatial curvatures, as illustrated in the sixth section. In addition, eigenvalues of the Hessian matrix and Hessian projections on certain surfaces also form measures of how curved the isocontours are.
[00105] Curvature measures the degree to which an object, a surface, for example, deviates from being flat. However, specific definitions depend on the context. For example, the curvature can be defined for a plane or a space curve, in which case it is a scalar. It can also be defined for a parameterized surface embedded in the three-dimensional space, in which case it is a tensor. In order to discover attributes related to stratigraphy, two new types of curvature are defined here. The first is isocontour curvature in a two-dimensional seismic image. An arbitrary two-dimensional cross section through a seismic volume is treated as a map of a function f (xy). An isocontour is the curve of this cross section whose coclassifieds satisfy f (x, y) = constant. Its curvature is defined as
where all quantities are defined by components of the isocontour gradient vector or Hessian tensor. This curvature attribute can be particularly useful in highlighting geometric aspects of high curvature, such as endings, corners, and joins.
[00106] Another definition of isocontour curvature is based on a three-dimensional generalization. Seismic data is treated as a function of three independent variables f (x, y). All voxels that satisfy f (x, y) = constant constitute an isosurface, that is, a curved surface embedded in three-dimensional space. For a curved space, curvature is a tensor quantity that allows extraction of multiple scalar quantities related to curvature, such as the main curvature, medium curvature, and Gaussian curvature. For such a curved surface, a normal surface can be defined by normalizing the isocontour gradient n = Vf / | Vf |. Surface tangents can be defined using the condition t n = 0. It is observed that there is a finite number of tangents that satisfy this condition at any given point on the surface. And there are also many planes orthogonal to the surface defined by a tangent t and the normal n. The intersection of such a plane with the surface is a curve, the curvature of which is called the normal curvature. The maximum and minimum of all normal curvatures are designated as the main curvatures K1 and K2, where the Gaussian curvature is computed as the product K = K1 K2, while its algebraic mean obtains the mean curvature M = (K1 + K2) / two.
[00107] All these quantities are defined for each point in space, and can be expressed compactly as an S matrix known as the Format Operator. To construct a convenient embodiment of S, it is observed that the space of all tangents of the given point can be expressed with the eigenvectors of a matrix T obtained by subtracting the external product from the normal with the identity matrix itself, T = (I - nm). T effectively projects any three-dimensional vector onto the tangent plane. Thus, one can measure the change from the restricted gradient to that plane by multiplying the Hessian matrix by T or, symbolically, H-H T, where ■ indicates the interior, or matrix, product. Finally, S is obtained by normalizing the Hessian: S = H ’/ | Vf | 2. The main curvatures are the eigenvalues of S, dividing the S feature by two fields, the mean curvature, M = feature (S) / 2, and the determinant det (S) = K define the Gaussian curvature.
[00108] All these curvature measurements can be treated as seismic attributes and can be used to detect horizon terminations and junctions that define aspects of geological importance, such as faults, truncations, pinchouts, and other geometric relationships.
[00109] The isocontour attributes are in the form of vectors and second order tensors. The projections of these attributes in particular directions and surfaces can characterize the seismic volume better than any of its individual components. The projection of the gradient over any direction is simply the point-product between the gradient vector and a unidirectional unit vector. Such a projection defines the directional derivative of an isocontour function along a particular direction. The larger the projection (or directional derivative) is, the steeper the change in amplitude along that direction. Similarly, the Hessian projection on a unit directional vector d is the second order derivative of an isocontour function along that direction and can be computed as fdd = d H d, where ■ indicates the interior, or point, product. For Hessian, projections of particular interest are along the gradient direction, direction of inclination, or direction orthogonal to the plane traversed by both gradient and inclination.
[00110] The isocontour slope volume gradient and its projections also contain useful information from the subsurface structure. The slope gradient quantifies slope direction changes in the original volume. The slope gradient projection along the slope direction is a measure of the rate of slope change. This helps to detect curved aspects in the reflectors, such as ridges, hills, cavities and valleys. The slope gradient projection on a vector that is normal to the slope direction defines a measure of the reflector's convergence or divergence rate.
[00111] Although these isocontour attributes (gradient, slope, orientation, Hessian, curvature, and projections) can be used as local attributes, the results can be affected by artifacts in areas contaminated with noise or areas with high complexity. Planing reduces such artifacts and can be carried out, for example, by convolution with a large low-pass filter or application of a medium filter with a large window. Preferably, however, planing is carried out in a structure-oriented manner over greater distances along the fabric to avoid mixing across the strata. By applying a long-distance, structure-oriented filter, local isocontour attributes can be converted into geological attributes.
[00112] 1c. Convergence
[00113] Another new geological attribute of the present invention is convergence, which detects where seismic reflections converge to form a tendency for regional thinning of reflection packages, highlighting pinchouts, wedges, or stratigraphic onlaps and downlaps. The convergence attribute can be used to emphasize stratigraphic aspects to an interpreter or as input to an assisted or automated standard recognition system, such as the seismic hydrocarbon system analysis system of the present invention. Two methods of computing convergence are: by using Hessian or by using a gradient gradient in normal tissue directions or inclination. Depending on the neighborhood used to compute the Hessian or gradient, long-distance structure-oriented flattening along the fabric can be employed using integral line convolution, a diffusion filter, a skeleton-guided filter, or domain filtering deity. Each of these uses the original seismic data to guide the filter through the reflectors.
[00114] Another method to compute the convergence attribute is to first measure the thickness of the local reflector, then estimate a derivative of local lateral thickness and, finally, perform a lateral planing operation along the reflections to obtain regional convergence trends. . If trends are estimated in single two-dimensional cross sections, then convergence can occur on the left or right or, alternatively, be called convergence and divergence along a reference direction, for example, left to right. If trends are estimated from orthogonal slices extracted from a 3D data set, then convergence can be associated with an orientation and magnitude. In addition, trends can also be estimated in a volumetric sense, using three-dimensional analysis windows instead of multiple orthogonal two-dimensional analyzes.
[00115] Methods for estimating local reflector thickness include: circuit duration, the distance between juxtaposed ends or zero crossings; instantaneous frequency, or the time difference between two juxtaposed surfaces of the seismic skeleton. A continuous attribute, such as instantaneous frequency, may need to be blocked. Any such thickness attribute can be differentiated laterally to assess the rate of change in local thickness. The rate of change in local thickness can also be estimated by the divergence of a vector field, for example, reflection normals, or by the divergence of a sensor field, for example, reflection tangents.
[00116] To obtain regional convergence trends, derivatives of local thickness are flattened with a structure-oriented long-scale filter operator along the fabric, using integral line convolution, a diffusion filter, a skeleton-guided filter, age domain filtering, or any other method that uses the original seismic data to guide the filter through the reflectors.
[00117] For computational efficiency, labels or horizon indicators can be coded with derivatives.
[00118] Figure 7 shows the convergence attributes based on the circuit duration, lateral derivative, and employing integral line convolution to increase long distance convergence trends. By performing this procedure both in lines and in crossed lines, a three-dimensional convergence attribute with magnitude and orientation can be constructed (Fig. 8).
[00119] 1.d Confluence
[00120] Another new geological attribute of the present invention is called confluence. Seismic reflections are treated as a network of paths carrying traffic, and some of these paths are congested because they are linked with many others. The confluence measures this congestion, estimating how many routes pass through any given voxel. In terms of seismic stratigraphy, branching and joining reflections produces terminations in the forms of downlaps, onlaps, toplaps, and truncations, as shown in Fig. 9. All of these terminations are potentially associated with non-conformities; downlaps can also be associated with surfaces of downlaps, that is, flood or transgressive surfaces.
[00121] A family of methods used to compute confluence is based on the density of the flow lines specified by the reflection tangents. Flow lines are computed by integrating the tangent field or simply following a discretized tangent direction chain. For example, starting with the leftmost line, a new flow line can be started and propagated to the right. The new flow lines are initiated in regularly spaced depth increments or two-way travel time, or at specific events, such as peaks (maximums), valleys (minimums), and / or zero crossings. Samples in a flow line, and preferably in a small area around the flow line, are marked to belong to that particular flow line. Once all the flow lines of the first trace are started and marked, the algorithm proceeds to the neighboring trace on the right and starts to start and mark additional flow lines, in regular increments in areas that have not been marked up to date or in events that have not yet been marked. Flow lines can be drawn and marked on the left or right; preferably, the flow lines are computed and marked both to the left and to the right of the current trace. Once all unmarked areas or events have been used to start a flow line and thus be marked, the process is repeated with the next stroke.
[00122] When the entire sample or event has been marked at least once, the confluence is determined by computing how many times each sample has been marked. Preferably, this calculation is performed simultaneously with the sample / event marking, simply by adding associated counters at the sample locations. Although these steps can be performed from left to right, they are preferably repeated from right to left as well and then added to the previous result, ensuring symmetry. The marking of samples in a region around a flow line can be done with a constant indicator or with values that vary depending on your distance from the flow line. Variable indicator weights, for example, could look like a triangle or a Gaussian. Disregarding the problems of when and where to start new flow lines, one could simply mark and count the flow lines without marking any regions and then, then, perform a planing operation, for example, with a trolley, a triangle or a Gaussian filter.
[00123] Instead of drawing flow lines from all directions, from left to right (and vice versa), flow lines can be computed for only a finite distance from the initiation point. This computation can be performed by integral convolution of the line or, preferably, a variation thereof that simply increases the counters of the samples touched by the stencil.
[00124] Generalizations of these embodiments based on the confluence flow line for three-dimensional data can be obtained by performing computations independently along the in-line and cross-line directions, or along a set of arbitrary directions. Alternatively, flow lines (curves) can be generalized to flow sheets (surfaces), or line integrals can be generalized to surface integrals.
[00125] Other forms of confluence can be obtained using seismic skeletons. The raw skeleton is represented by a directed event graph (peaks, valleys, and zero crossings: vertices) and a waveform correlation with events in neighboring features (edges). From a particular vertex (event), the graph can be crossed to the left and / or to the right, following the edges (correlations). Starting the chart crossings of different vertices, it is possible to count how many times any given vertex obtained passing through. Every vertex is used to start a crossing or, preferably, only vertices that have not been passed through previously are used to start a new crossing. In the preferred case, first, the entire vertex will start a crossing, but progressively, vertices will have passed through at least once and the beginning of new crossings will become increasingly rare. This particular embodiment of crossing confluence graphs can also be based on morphological skeletons.
[00126] Two other forms of confluence are based on the topological skeleton. First, one can count how many times different surfaces overlap or overlap any given surface. If a surface below another particular surface ends, then a new surface must exist 3M57 below the original (unfinished) surface. The score of how many different surfaces are above and below a given surface, therefore, approximates the number of endings next to it. By their nature, large surfaces tend to have high scores. Normalizing scores by surface size, however, tends to designate high relative scores for the smallest surfaces, because a small number of scores is divided by a small area or a small number of samples forms the surface. Thus, scores need to be normalized. One method of normalization involves using the increased surface area of some capacity, for example, one half. A preferred normalization, however, is based on analyzing the significance of finding a number of terminations on a surface of a particular size. For a small surface, for example, there is a chance of finding a large number of terminations in relation to the surface size and, therefore, this score cannot be statistically significant. For a large surface, however, even a short-ended score can be above average in a statistically significant way. The probability of finishing a surface can be estimated by averaging the relative individual finishing scores, some weighted version, or the total number of surfaces divided by the total area of these surfaces. Once an expected termination probability (or density) is calculated, a binomial test can then be used to compute the statistical significance of a deviation from the number of terminations expected for a surface of a given size. For large surfaces, the binomial distribution is well approximated by convenient continuous distributions, and this can be used as the basis for alternative tests that are much faster to compute, that is, Pearson's square test and the G-test. However, for small samples, these approximations are exhausted and there is no alternative to the binomial test.
[00127] Another form of realization of confluence derived from topological skeletonization is based on the age mapping volume associated with the surfaces. High confluence areas have a high density of flow lines or flow sheets. An isocontour or an isosurface of the age mapping volume is similar to a flow line or slide derived from reflection tangents. Thus, a high density of isosurfaces indicates an area of high confluence. A preferred alternative for building isosurfaces and estimating their density is to compute a gradient or vertical derivative of the age mapping volume. Areas with high gradients or high derivatives have a high confluence.
[00128] Any forms of confluence can also be homogenized by applying a long distance planer, for example, a line convolution integral.
[00129] 1.e Spill and Closing Points
[00130] Another set of geological attributes, described in US Patent Application Publication No. 2010/0149917, “Method for Geophyisical and Geological Interpretation of Seismic Volumes in Depth, Time, and Age”, by Imhof et al, is the envelope and spill points. In traditional interpretation practices, a horizon is analyzed to determine the existence of closed contours involving a high topographic site, forming a closure, which could trap hydrocarbons. Note that the closure could be implicit, for example, where the contour lines end in a fault or against a saline dome and are implicitly allowed to follow these limits. For each closed contour, wraps specify the area contained within. Thus, any location on a surface can be analyzed to determine whether or not it is located in a closed contour surrounding a high site and, if so, the enclosed area can be computed. For each elevated site, its maximum closure specifies the maximum extent of the potential hydrocarbon trap.
[00131] Closures and wraps can be determined by any single horizon or by any set of horizons, for example, surfaces mapped by traditional means or by automatic skeletonization. By estimating the envelope for each surface in a set of age volume data, or preferably, the depth volume, a envelope volume can be computed that assigns each sample a value of zero, if not starting from a contour of closed depth that circulates a high site. Otherwise, the area, from the surrounding area, is assigned to the sample site.
[00132] Using the age volume, a surface can be constructed by selecting an age, or a location, or some other criteria. This surface can be analyzed for elevated sites, contours that circulate the elevated sites, and their areas. Preferably, however, the depth volume is employed. Each horizontal slice in a depth volume represents a surface at some age. In fact, a volume of depth is nothing more than a pile of surfaces classified by age. For each slice (or surface of some age), its values correspond to the depth and, thus, each slice constitutes a depth map. Each slice can be analyzed for high sites, closed contours, and involved areas that allow the computation of a wrap volume in the age domain. If desired, the entire wrapping volume or part thereof, a slice, for example, can easily be transformed into a depth domain.
[00133] An extension of closure or wrap volumes are spill points and spill points volumes. The spill points are a location close to the maximum closure contour where the contours are broken and, thus, the potential trap leaks (Fig. 1). The determination of spill points allows the creation of a volume of spill points and an examination of how different potential traps spill and feed among themselves. The locations and number of spill points can be used to risk a prospect or to guide an investigation as to the regions where hydrocarbons have leaked from and accumulated in. Although spill points can be identified from age volumes by extracting isoity surfaces and examining them, spill points and spill point volumes are preferably generated from depth volumes that correspond to stacks of different depth maps. ages (ie stacks of depth maps for different horizons). If necessary, spill points and spill points volumes, determined by depth volumes, can be easily transformed into a depth domain, using the depth volume as a look-up table or by interpolation.
[00134] 1.f Phase Waste
[00135] Other geological attributes may be formed from instantaneous phase anomalies, the points of which are located outside where the seismic ripple divides (PCT Patent Application Publication WO 2009/137150 Al, “Method For Geophysical And Stratigraphic Interpretation Using Waveform Anomalies ”, By Imhof). The instantaneous phase anomaly attribute is a different attribute indicating whether or not a small wave division occurs at a sample site. Traditional convolution with a long pulse filter allows the computation of a local density of ripple divisions. The long-distance structure-oriented planing, for example, with integral line convolution, allows the computation of the density of ripple divisions, which is consistent with the underlying seismic tissue. Preferably, ripple dividing circuits and strands are first classified based on their shape and orientation to suppress ripple divisions caused by noise or artifacts.
[00136] The phase residues occur in places where waveforms divide, which implies that an additional reflection event begins or that a reflection is incorporated in another. In other words, a termination has occurred. The terminations are often classified as onlap, downlap, toplap and truncation (Fig. 9), depending on the relationships between the reflections. Thus, phase residues can be classified as onlaps, downlaps, toplaps, and truncations. As an alternative to phase residues, terminations can also be detected and classified as seismic skeletons. Unless the slope information is used, downlaps and onlaps cannot be distinguished and may need to be grouped together. Without slope, toplap and truncation information they cannot be distinguished and need to be grouped together. The terminations and their score can be affixed to surfaces or confluence flow lines. Preferably, the terminations and their scores are converted into density or geological attributes by flattening. The particularly useful geological attributes based on terminations are termination densities, such as onlap / downlap density, truncation / toplap density, or the difference between onlap / downlap density and the truncation / toplap density that indicates which duct dominates.
[00137] 1.g Texture
[00138] Other geological attributes are based on seismic texture. US Patent No. 6,438,493 BI, “Method for seismic facies interpretation using textural analysis and neural networks”, from West and May, describes a method for the identification of seismic facies, based on textile attributes computed with a co-occurrence matrix of gray level (GLCM). Local texture attributes derived from GLCM include, but are not restricted to: texture of homogeneity, inertia (also known as the element difference moment or contrast), entropy, and energy (also known as uniformity). Using long distance planing along the fabric, these attributes can be converted into regional geological attributes related to the texture of seismic facies. GLCM-based texture attributes can further be generalized into three-dimensional texture attributes (for example, U.S. Patent No. 6,226,596 B1), “Method for analyzing and classifying three dimensional seismic information”, by Gao).
[00139] Another texture attribute (PCT Patent Application Publication WO 2009/011735 "Geologic Features From Curvelet Based Seismic Attributes", by Neelamani and Converse) identifies stratigraphic aspects of seismic data, collecting a data transformation from a curve small. From this small curve representation, selected geophysical data attributes and their interdependencies are extracted and used to identify geological aspects. Using long-distance planing along the fabric, these attributes can be converted into regional geological attributes related to texture and seismic facies.
[00140] Other texture attributes are based on multidimensional Fourier measures provided with a window (PCT Patent Application Publication WO 2010/053618, “Method for Seismic Interpretation Using Seismic Texture Attributes”, by Imhof). One particular measure is regularity, a texture attribute that measures how banded (or regular) seismic data appears to be. The free areas, through reflections, exhibit a high degree of regularity, while areas of noise with disorganized reflections exhibit low regularity. With small analysis windows, regularity is a measure of discontinuity. With large windows of analysis, regularity acts as a measure of disorder. Thus, regularity can also be called chaos of multipath discontinuity. Using a long distance flattening across the fabric, these multidimensional Fourier attributes provided with windows can be converted into regional geological attributes related to texture and seismic facies. For example, crossing regional reflections, commonly called railroad tracks, often correspond to the transgressive surfaces (or flood surfaces) that often form fences. Thus, the regularity long-distance flattening (or its reverse, chaos) creates a geological attribute related to hydrocarbon seals and, due to shales, form both seals and act as matrix rocks, hydrocarbon matrix rocks.
[00141] Randen and Sonneland (“Atlas of 3D Seismic Attributes”, in Mathematical Methods and Modeling in Hydrocarbon Exploration and Production, Iske and Randen (editors), Springer, pp. 23-46 (2005)) present a summary of the seismic attributes additional three-dimensional features that characterize the seismic texture or earthquake-stratigraphic aspects.
[00142] 1.h WPCA anomalies
[00143] Another family of attributes relevant to this invention is described in PCT Patent Application Publication WO 2010/056424 "Windowed Statistical Analysis for Anomaly Detection in Geophysical Datasets", by Kumaran et al. These attributes highlight locations in one or multiple data sets where the seismic data is statistically anomalous compared to other locations. In addition, these attributes generate a base of vector patterns ranging from the common majority to the anomalous majority. Designing the data in a linear combination of these patterns emphasizes some patterns, suppressing others. For example, the ubiquitous seismic bandage can be suppressed. Using long-distance planing along the fabric, the seismic filtrate can be converted into regional geological attributes. Another application is the local decomposition of the data in these patterns and the determination of the dominant pattern, whose score of effects of the seismic samples is based on the patterns of the surrounding sample. Using a medium or long distance filter along the fabric, the local class can be converted to a regional one.
[00144] 1.i Other Attributes
[00145] Another type of attribute is based on the inversion of multi-displaced seismic data and the observed amplitude versus displacement behavior that allows the prediction of porosity and clay content, and allows the lithofacies score, (for example, US Patent No. 7,424 367 B2, Minutes of the 2008 conference on “Method for predicting lithology and porosity from seismic reflection data”, by Saltzer et al., “Seismic Rock-Property Inversion and Lithofacies Prediction at Erha Field, Nigeria”, Xu et al, Nigerian Association of Petroleum Explorationists (NAPE); “Lithofacies Prediction in Deep Water Water Reservoirs”, Oppert et al, Society of Exploration Geophysicists, Expanded Abstracts, 1708-1711, (2006)). Using long distance planing along the fabric, these attributes can be converted into regional geological attributes related to lithofacies, clay content, and porosity.
[00146] European Patent No. EP 1110103 Bl "Method Of Seismic Signal Processing", tode Meldahl et al., Describes a method for generating attributes that allow the detection of potential gas chimneys. A gas stack is a vertical seismic response disturbance of gas filtration that degrades seismic data due to acquisition and processing limitations. The resulting chimney hub attribute highlights vertical disturbances of seismic signals that are often associated with gas chimneys. It reveals information about the history of hydrocarbons and fluid flow. In other words, the chimney cube can reveal where hydrocarbons originated, how they migrated into a prospect and how they move from these prospects. As such, a chimney cube can be seen as a new indicator tool for indirect hydrocarbons.
[00147] In addition, Loseth et al. (“Hydrocarbon leakage interpreted on seismic data,“ Marine and Petroleum Geology 26 (7), 1304-1319, (2009)) provides a review of seismic hydrocarbon leakage interpretation.
[00148] Finally, the seismic attributes can be related to the properties of the subsurface that can be used to simulate transport phenomena, such as heat and temperature flow, which affect hydrocarbon maturation, or permeability and fluid flow, which affect hydrocarbon migration (PCT Patent Application Publication WO 2009/137228 A2, “Transport Property Data Calculated From Derivative Seismic Rock Property Data For Transport Modeling”, by Oppert et al).
[00149] 1 .J Combination Attributes I Play Attributes
[00150] In some embodiments of the present inventive method, seismic data is examined for one or multiple specific plays, rather than elements of a more generic hydrocarbon system. Examples could include: anticline plays (Fig. 10), normal-fault plays (Fig. 11), dome-salt flank plays (Fig. 12), shoe cord channel plays, or others, as illustrated in Figs. 13 to 16. In these cases, attributes, detectors or workflows are necessary to distinguish specific aspects, such as faults, salt, or channels.
[00151] PCT Patent Application Publication WO 2009/082545 “Detection of Features in Seismic Images”, by Kumaran and Wang, describes such a method for detecting channels or flaws in seismic data. For fault detection, edges are identified in flattened seismic images and the edge intensities are integrated in multiple directions, for example, using Radon transformation to detect the presence and orientation of fault lines. For channels, edges are detected and converted to smooth curves to identify channel edges. Sets of smooth parallel curves are then examined to find pairs of curves that correspond to the left and right channel edges, thus defining the channel.
[00152] US Patent No. 7,203,342 B2, “Image Feature Extraction,” by Pedersen, describes another such method designed to extract flaws from seismic attribute data, although it can also be used for other line or surface extraction problems as well . The method is based on antenna tracking, where antennas or numerical agents roam through a discontinuity or edge detection volume, slowly connecting parts close to aligned edges.
[00153] A final example is the detection of saline bodies using attributes of regularity or chaos described in PCT Patent Application Publication WO 2010/053618 "Method for Seismic Interpretation Using Seismic Texture Attributes", by Imhof; or Randen and Sonneland, “Atlas of 3D Seismic Attributes” (in Mathematical Methods and Modeling in Hydrocarbon Exploration and Production, Iske and Randen (editors), Springer, pp. 23-46 (2005)).
[00154] These are just three exemplary methods for identifying specific aspects. Many others have been described and are well known to practitioners of the art.
[00155] Partition
[00156] In order to detect the simultaneous presence of multiple proximal elements of the hydrocarbon system or multiple play elements, the data volume can optionally be divided into at least one segment for analysis, perhaps in combination with a background segment that do not be analyzed. Typical modes of partition are the analysis of: individual voxels, small blocks, tissue-aligned blocks, layers, or groups of contiguous voxels. The partition does not have to be mutually exclusive. Individual divisions can overlap.
[00157] The simplest partition is the voxel-by-voxel analysis, but the results can be done by patches, because rarely all elements are recognized in the same voxel and it is unlikely to find an extended region of contiguous voxels with all elements gifts. Such techniques, such as flattening detected attributes or elements extended in surrounding voxels, can be employed in step 5 of the present inventive method (Fig. 2), evaluation, to create larger contiguous prospects. Partitioning into voxels is a preferred method of division, as the data is often readily represented as voxels. For this reason, the partition step is considered optional. In addition, partitions (and thus voxels) can be aggregated in larger regions during the evaluation stage. Thus, for the purposes of teaching the inventive method, if no division has been explicitly performed, then each voxel is considered to form its own division.
[00158] Another partition scheme involves interrupting the volume of data in regular Cartesian blocks or bricks, for example, samples of size 20x20x20, where the expectation is that some or all elements are present within a prospective block.
[00159] Cartesian bricks will cut through the layers and fabric. An alternative to regular Cartesian bricks or blocks is to align the blocks with the fabric. In this scheme, there will be differences in size and shape between the divisions as they conform to the fabric. A particular way of generating such a division is to use the fabric itself to define a thick layer structure. Specifically, for example, a selection of surface parts created by seismic skeletonization could be extended vertically to create bodies or segments.
[00160] Another preferred method for division is based on one or multiple seismic attributes as generated, for example, in step 2. Partitions are created by attribute thresholds, followed by connected component analysis or similar to generate contiguous regions embedded in one second plan. This process is also called seed detection (multivolume). A preferred attribute for controlling the split is the protrusion, an attribute that highlights locations in one or multiple data sets where the seismic data is statistically anomalous compared to other locations. The protruding attributes are described in PCT Patent Application Publication WO 2010/056424 "Windowed Statistical Analysis for Anomaly Detection in Geophysical Datasets", by Kumaran et al. The boss is a generic name for an attribute that highlights statistical anomalies in data. Kumaran et al. describes “inverse covariance”, “(WPCA) residue”, or “(WPCA) anomaly,” terminology for specific boss embodiments. Additional overhang attributes can be computed using Kumanan methods. The projection can be taken as a threshold to perform the analysis only of contiguous anomalous regions.
[00161] An alternative method of using attributes for division of control is to first perform a different method of division, for example, Cartesian blocks or aligned tissue bodies and then preserve only divisions where an attribute computed for each division exceeds a value absolute or a relative value, or satisfy some prescribed condition, for example, where the maximum salience is in the top quintile or where the average regularity is within the range between 0.3 and 0.6.
[00162] A final example of a division method involves subdividing the volume of data into small regions that are classified by an attribute (s), for example, boss. For many embodiments of the present invention, the order in which someone assesses the prospects of the different divisions or individual voxels does not matter. This last example demonstrates the use of an attribute, such as overhang (or size, another attribute, or a combination of other attributes) to determine the order in which divisions are analyzed. Using such a prioritization may not be necessary to analyze all divided regions. This embodiment of the present inventive method begins the evaluation of the hydrocarbon system (step 5) with the most salient regions and continues with progressively fewer protruding regions until a prescribed number of regions has been analyzed, a prescribed number of prospects have been found, or a prescribed time allowed for analysis has been exceeded.
[00163] Element Scores
[00164] The score can include normalization, conditioning, combination, or 42157 escalation. For some elements of the play system or hydrocarbons, this step admits at least one attribute per element available to form a score for each partition (or in the simplest case, for each voxel) that expresses the probability or expectation that a particular element will be contained in a particular division. The score is supposed to be an optional step, because it can be performed as a separate explicit step 4, as shown in the flowchart of Fig. 2. In some embodiments of the invention, however, it can be realism in combination with the formation of an attribute (step 2) or prospectus evaluation (step 5).
[00165] In the simplest form of the inventive method, an attribute is directly used as a score to indicate whether or not a particular element is present in any given location or not. Different attributes, however, can have variations for their values and are often advantageous for normalizing their values, for example, between 0 and 1 or 0 to 255, for easier comparisons between attributes or to facilitate probabilistic interpretation. Normalization is just a linear transformation of attribute values.
[00166] An alternative to normalization is the calibration or application of a nonlinear transformation of attribute values that is triggered by data, for example, histogram equalization or histogram transformation, or triggered by a model that expresses how a related attribute up with the probability for a given element to exist. Such a model can be based on theory, measurements, prior or contextual knowledge, experience, or intuition.
[00167] Seismic attributes are observations, measurements, or computations performed on seismic data. They can relate to elements of the hydrocarbon system or play elements, however, they often do not really measure these elements. In addition, they may not be unique. Different attributes can refer to the same element. The same attribute can be related to multiple elements, while also affected by noise, acquisition, data processing, and the algorithm and parameters used to generate the attribute. The same type of attribute can be computed with different algorithms. The combination of attributes allows the formation of a score that indicates the probability or expectation for the presence of an element in a given location.
[00168] If for a given element, no direct attribute has been computed that defines a score, then the score may need to be defined indirectly, using a substitute attribute or score, a heuristic, a concept, or a previous expectation (such as as a constant value of 0.1 that simply indicates a 10% chance for the element). The reservoir, for example, can be directly indicated by an estimate of sand-shale ratio and / or an estimate of porosity. If no such indicator attribute exists, then a conceptual substitute, derived, for example, from stratigraphic sequence concepts, may need to be used. Such a substitute attribute could be a combination of a lowstand basin fan that overlaps a non-conformity (a sequence limit) as shown in Fig. 17. Lowstand basin fans often exhibit good porosity and permeability and thus have the potential to form the reservoir element. Raising the relative sea level will cover that fan with the fingers of the lowstand wedge, probably to consist of thinner, less permeable matter, which is covered by a transgressive surface. Further relative sea level rise will bury the lowstand below the tracts of transgressive and hightstand systems. In distant locations, the fan may appear to be sandwiched between a "non-conformity" and a "transgressive" surface. Thus, a rock body between a non-conformity and a transgressive surface, located on the deep side of said non-conformity, could be a lowstand fan and would therefore be designated a high reservoir score.
[00169] Another aspect of score designation is the escalation of segments or partitions. Many attributes are defined at each voxel location. Others may only be available on surfaces or in selected locations. A particular partition can correspond to a voxel or to contiguous sets of voxels. In order to assign a score to each partition (or in this case each voxel), attributes may need to be interpolated or reduced to obtain a single score per partition. For a partition that contains multiple samples, a representative score can be found by computing the average, average number, or mode; the application of a voting procedure; or selecting the minimum, maximum, or random sample value.
[00170] It is advantageous to increase the scores with confidence. The score represents a probability of finding a particular element in a given location. A score, however, does not indicate how credible this probability is. A place where a seismic attribute predicts high porosity can receive the same score as a place with a lowstand fan, as suggested by the proximity of a non-conformity and a transgressive surface. However, at the first location, there is a direct relative measurement of porosity and, thus, the presence of a reservoir. In the second location, the existence of a reservoir is made known by a conceptual model. Direct measurement inspires higher confidence than a prediction based on a conceptual model. This difference is expressed in the confidence value associated with a score.
[00171] Prospectus Evaluation
[00172] Using the scores for the elements of the play or hydrocarbon system allows the evaluation of partition (or voxel) prospects (step 5) and the identification of their deficiencies, that is, elements that are weakly expressed, deficient or unresolved .
[00173] Many of the described geological attributes represent trends and thus have a relatively weak resolution. In some embodiments of the inventive system, the needs for the spatial arrangement of elements, for example, sealing rock above the reservoir, are relaxed and replaced by a test for the presence of all elements within a single partition. This test for the presence of all elements is called a logical test provided with doors, because a division has to pass the test by all elements in order to be judged prospective. The logical approach provided with matching element doors, therefore, is an “all or nothing” approach. Satisfying an element frequently means exceeding a score threshold that is typically specified by the interpreter or coded in the system. Unsatisfactory elements can be indicated for another examination because they constitute the weak connections of a partition. The logic provided with doors is a preferred embodiment of prospect evaluation.
[00174] A more gradual prospective test is a vote that simply counts how many elements are present in a given voxel or partition. The more elements present, the more a division is considered prospective. An element is present or absent based on a comparison of its score against a threshold specified by the interpreter or coded in the system. Missing or weakly expressed elements can be flagged. Voting is another preferred form of conducting the prospectus assessment. A variation in voting is weighted voting in which some elements receive a higher weight or more votes than others. Weighted voting could be used, for example, to emphasize elements that are detectable with the highest confidence.
[00175] Combinations of logic provided with voting doors are possible, for example, requiring that a specified number of elements be present instead of requiring the presence of all elements. Another example is the necessary presence of some specified elements and the desired presence of others.
[00176] The prospectus assessment may include aspects of step 4, score. An example is the combination of different attributes, for example, attributes computed with different algorithms or parameterizations, which refer to the same element (s). Instead of first combining the attributes in scores for individual elements and then evaluating the prospect based on the scores, the prospect of partition can be evaluated by different attributes, for example, by voting or counting how many threshold attributes the user or system-specified exceeded.
[00177] Rigid thresholds can be avoided by using a multi-valued logic, for example, fuzzy logic, which is derived from the fuzzy set theory to deal with reasoning, which is approximate rather than precise. The logic provided with gates is clear and binary, with association values 0 and 1 representing below the threshold (missing element) and above the threshold (present element). Fuzzy logic scores, instead, have association values ranging from 0 to 1 and represent the degree of truth of an assertion. Both the degree of truth and the degree of probability can vary between 0 and 1 and therefore can appear similar. However, they are conceptually distinct. Truth represents association in loosely defined sets, not probabilities of any event or condition as in probability theory. Obtaining, for example, a stratigraphic layer that contains 70% of sediment and 30% of shale. We can consider two concepts: reservoir and sealing. The meaning of each of them can be represented by a certain nebulous set. One could define the formation as being 0.7 reservoir and 0.3 sealing. It is observed that the concept of reservoir would be subjective and, thus, would depend on the observer or designer. Another designer could equally and correctly build a set of associated functions in which the formation would be considered a reservoir if the sediment part exceeded 50%. Fuzzy logic uses degrees of truth as a mathematical model of the uncertainty phenomenon, while probability is a mathematical model of randomness. A probabilistic scenario would first define a scalar variable for the sediment fraction and then define conditional distributions describing the likelihood that someone would call the formation of a given reservoir of a specific sediment fraction. It is observed that the conditioning can be obtained by having a specific observer randomly selecting the label for the layer, a distribution through deterministic observers, or both. Consequently, probability has nothing in common with diffusibility; these are simply different concepts that superficially look similar because they use the same range of real numbers between 0 and 1. Furthermore, the confusion arises because properties of random variables are analogous to properties of binary logical states, and theorems, such as the De Morgan's, which refers to logical operators “e” and “or” in terms of each other via negation, have double applicability.
[00178] An alternative to fuzzy logic is Bayesian logic, which is based on the Bayesian probability theory, which makes it possible to reason with uncertain statements. To assess the probability of a hypothesis or configuration, a human or machine interpreter specifies some previous probability, which is then updated in the light of new relevant data. The Bayesian interpretation provides a standard set of procedures and formulas for performing this calculation. One method of integrating the needs of a particular configuration and determining its likelihood is to use a Bayesian Belief Network (BBN). A Bayesian network is a model based on a probabilistic graph that represents a set of random variables and their conditional dependencies, via a directed acyclic graph (DAG). This graph indicates the structure of conditional independence between random variables representing the different elements and their spatial arrangements. For example, a Bayesian network could represent the probabilistic relationships between elements of play and types of play. Given certain elements of play, the network could be used to compute the probability of the presence of various types of play. Formally, Bayesian networks are directed acyclic graphs whose nodes represent random variables in the Bayesian sense: they can be observable quantities, latent variables, unknown parameters or hypotheses. Borders represent conditional dependencies; nodes that are not connected represent variables that are conditionally independent of each other. Figure 18 presents an exemplary network graph for the entire hydrocarbon system. The scores for the reservoir and sealing are combined by analyzing their overlap, which indicates whether the reservoir is covered by the sealing rock. Given the local geometry, the reservoir / sealing system is called a trap score. Preferably, the trape is filled with hydrocarbons and thus the trape score is then combined with scores for the presence of a source and indications of at least the potential for migration paths that go from the source to the trape. Ideally, there is a direct indication for hydrocarbons (DHI) in the seismic data as well. Therefore, the scores for trape, source, migration, and DHI are all combined to produce an accumulation score. All accumulation scores can now be weighted by size and confidence to risk and classify potential hydrocarbon targets.
[00179] Even when employing geological attributes that represent trends, partitions can be too small to contain all elements, for example, when dividing into individual voxels. Two approaches to testing coexistence or placement within small partitions or between voxels are relative spatial changes and the extent of regions of influence. Scoring Change converts a test spatially for a given spatial relationship to a placement test. For example, the sealing score can be changed downwardly relative to the reservoir score to examine the existence of sealing rock through the reservoir. Without the application of such a change, it is necessary to test for sealing on the reservoir. With the application of such a change, one can examine the presence of a seal and reservoir, for example, with a logic provided with doors. Rather than using a single change, it can be advantageous to assess the coexistence of two changes for a range of changes, giving preference to minor changes; for example, designating a higher confidence for smaller changes to indicate the closer proximity of the required elements or reducing prospective as a function of distance. Side changes can be used to detect elements of play, such as faults, near the reservoir. Instead of applying vertical or lateral changes, changes can be applied across the fabric or perpendicular to the fabric. In addition, they can also be applied in arbitrary directions.
[00180] Change is inefficient if many different variations and orientations or directions need to be examined. A preferred alternative for the change is to extend the region of influence to raw scores or thresholds, for example, by convolution or morphological dilation. In the first case, scores are impaired or extended along specified directions, for example, by convolution with a tapered directional filter that decays with the band increasing, encoding the decrease in confidence. In a particular embodiment of the extension of the region of particular influence by convolution, the result of the convolution is added to the original score to form an updated score that is used for prospect evaluation. The extension direction is specified by an interpreter or determined by the fabric.
[00181] Gross scores or thresholds can be extended by directional morphological dilation that drags high scores along a specified direction in areas with lower scores. The direction can be specified by the interpreter, be coded in the system, or be derived from the fabric. Morphological operations can be applied to raw scores or arbitrary thresholds and partitions.
[00182] Figure 19 shows a schematic application of the inventive method for four partitions or potential target areas. Assuming that neither the source score nor the migration score can be determined by the data provided that do not direct hydrocarbon indications being observed, and that the remaining elements are independent, the prospectivity for each target could be computed by multiplying confidence with size and average of the reservoir, sealing rocks, and trap scores. Table 1 shows an example of this scoring process.

[00183] Table 1: exemplary classification for the targets in Fig. 19, of size, confidence, and the average scores for the reservoir, sealing rock and trap.
[00184] Instead of representing the scores for each element with unique values, as used in the example in Table 1, the scores can be represented by distributions that capture the measured scores and their uncertainties and / or confidence. Highly certain scores associated with high confidence have a peak distribution with less certain scores having a wide distribution. The scores that are needed, but not measured by any attribute, will have an even distribution. In some cases, distributions can be propagated and combined through the system (Fig. 18). The scores could also be integrated into a Monte Carlo model in which individual scores are randomly taken from the corresponding distributions and propagated through the system. Repeating these steps often allows for the formation of subsequent distributions.
[00185] Analysis and Visualization
[00186] Analysis and visualization (step 6) is an optional step that can be combined with step 5, prospectus evaluation. Analysis methods include: combining neighboring partitions (repartition), classification of prospects, validation of primary prospects, analysis of secondary prospects, and weaker link analysis.
[00187] Neighboring partitions can be combined to form larger contiguous partitions. A first example is the combination of voxel-single partitions in larger contiguous bodies, for example, by connected component analysis or multivolume seed detection.
[00188] The invention can use millions of voxels or tens to millions of partitions that contain at least some elements of the hydrocarbon system or at least some elements of play. Preferably, the prospectuses in which all the elements are present are classified in order to facilitate the validation of the prospectus. Preferably, partitions with at least some elements are classified in order to confirm or refute their faults. Classification can be performed, for example, by size or confidence.
[00189] The large partitions that contain all the elements are of primary interest for another evaluation, because they represent prospects, that is, areas in which the inventive system predicts the existence of hydrocarbons, and can finally be recommended for drilling. The justification for drilling a prospectus will be made by other traditional analyzes that are especially focused on the elements of least confidence. Also of interest are partitions that lack the least number of elements, for example, due to low confidence values. Such partitions can be analyzed with traditional methods to see if they can be reclassified as prospects or not. In this case, another investigation is preferably focused on the weaker connection of a potential hydrocarbon system or prospect associated with a partition.
[00190] Prospectivity, confidence, scores, flagged elements, and attributes are typically stored in memory or on the disk for further analysis and visualization.
[00191] Play selection
[00192] In some embodiments of the inventive method, data are examined not for generic hydrocarbon systems, consisting of reservoir, trap, sealing rock etc., but for specific plays (for example, a flank-saline play) that contains specific play elements. For a particular play to be considered, the interpreter needs to specify a configuration with the defining elements, their spatial relationships and appropriate geological attributes that relate to these specific elements.
[00193] Figures 10-16 represent some common plays. In all of these figures, the porous rocks (indicated at 101 in Fig. 10) are reservoirs that potentially contain hydrocarbons (indicated by dots). To prevent hydrocarbons from escaping vertically, a sealing rock is required (indicated by dashes). A trap or trap mechanism is necessary to prevent hydrocarbons from escaping laterally. Structural plays are created by deformation of geological strata, which include sealing rock and reservoir formations, in geometries (or structures) that allow the accumulation of hydrocarbons. Such resulting geometries, which involve reservoirs and sealing rocks, are dominated by folds (Fig. 10), faults (Fig. 11) or saline diapires (Fig. 12 - where crosses mark the saline diapirate). In stratigraphic plays, the trape geometry is formed by variations in the rocks (or stratigraphy) that are related to their deposition. An example of stratigraphic plays is represented in Fig. 13, where the formation of a reservoir, sand for example, tunes in the formation of sealing rock, shale for example. Other stratigraphic plays are related to the erosion of the reservoirs and the formation of non-conformities that are covered by sealing rock strata. Figure 14 represents a classic stratigraphic play, where the strata above the non-conformity (indicated by the full snaking line) provide the sealing rock for plunging reservoir formations that have been leveled by erosion. Figure 15 represents another stratigraphic play, in which the erosional relief in the porous reservoir formations is buried under the trap sealing rock formations. Figure 16 represents a configuration in which the reservoir is formed by a porous limestone, which is covered by a formation of sealing rock shale. Lateral variations in diagenetic processes preserved the porosity of the reservoir, however they obstructed the porous space in the upward diving direction, thus preventing the hydrocarbons from escaping laterally. The plays shown in Figs. 10-16 are by no means exhaustive, but simply represent examples. Many other cases have been described and are known to practitioners of the art (for example, Hydrocarbon Traps, K.T. Biddle and C.C. Wielchowsky. The Petroleum System - From Source to Trap, AAPG Memoir 60, pages 219 - 235 (1994)).
[00194] For each play, its elements of definition, its spatial relationships and appropriate attributes or geological scores need to be specified. For anticline play, stuck in a fold (Fig. 10), the minimum requirement is geometric in nature, that is, the presence of strata with anticline structure resembling an upside-down glass. Preferably, there is a reservoir formation with sufficient porosity and permeability to store and transmit fluids that is covered directly by a sealing rock formation, which is impermeable and forms a barrier to prevent these fluids from leaking. Ideally, there is a direct geophysical indication of hydrocarbons or at least some indication of source rocks and fluid migration paths into the trap.
[00195] For play stuck in failure (Fig. 11), the minimum requirements are also geometric in nature, that is, the existence of strata dipping upwards in a fault and the formation of a three-dimensionally closed reservoir compartment. If possible, one formation can be identified as constituting a reservoir, while another formation in close proximity above it is identifiable as a sealing rock. Ideally, there is a direct or indirect indication of hydrocarbons, a potential close source, or fluid migration paths.
[00196] A saline flank play (Fig. 12) is defined by strata diving upwards into a saline dome. A stratigraphic tuned play (Figf. 13) requires an upward dive formation that ends in a point. A stratigraphic non-conformity trap (Fig. 14) requires at least dive formations that are flattened by a non-conformity. The stratigraphic play of buried erosional relief (Fig. 15) requires at least one non-conformity with a three-dimensional arc shape. For all these configurations, additional criteria include: the presence of strata forming the reservoir covered by strata forming the sealing rock and the presence of a potential source, signs of fluid migration, or even direct indication of trapped hydrocarbons.
[00197] Schematic such as those shown in Figs. 10-16 are useful for developing, defining and transmitting configurations. Prospective plays can be found and analyzed in an ad-hoc manner with a specific workflow (or script) tailored or interactively by an interpreter. Some embodiments of the inventive system can be based on one or multiple workflows that are mostly independent.
[00198] Preferably, the specific plays and attributes used to characterize its elements are stored in a configuration catalog for reuse in similar situations.
[00199] Definition of a Configuration Catalog
[00200] Rather than defining plays and their configuration for each application of the invention again, it may be advantageous to create a catalog or library of plays and configurations for repeated use. The system interpreter can then select one or multiple types of play with their associated catalog settings. The making of such a catalog encourages the reuse of configurations, promotes reproducibility and facilitates validation. In addition, the catalog allows automated embodiments of the inventive system, which compares data or partitions against a set of potentially large configurations taken from the catalog. It might even be desirable to have different ways of performing the same play in the catalog, to allow variations in the quality of the data, or expression of the play. In this case, the user of the invention could look for a flank-saline play taking all the flank-saline forms of the catalog, perform the analysis using each one and use the best result of any given partition. The user could even browse the entire catalog through the partitions and then designate a type of play for each partition according to the probabilities.
[00201] For automated analysis, recognition, or comparison with a large set, or even the entire catalog, configurations are preferably represented in a formal manner, rather than ad-hoc, for each configuration. An example is a graphical representation for the configurations, in which vertices or nodes correspond to the required elements and express the preferred attributes, while the graphic borders mark their spatial relationships. Another representation of the catalog entries is in the form of a relational database.
[00202] The advantages of using a more formal representation are the reusability of the underlying components, rapid modification of existing configurations, rapid addition of new configurations and consistency between results. In addition, a more formal representation encourages the rigorous definition of spatial arrangement concepts, such as above or near. Finally, the system can potentially be scaled up in a recursive manner.
[00203] A less formal, but often preferred method of representing configurations in the catalog, is by assigning weights to geological attributes. Each play configuration is characterized by the weights assigned to the associated geological attributes or scores. Attributes or scores that are important for particular elements are given a high weight, while non-important ones are given a low weight, zero for example. The spatial relationships required between the elements are captured by regions of influence or changes designated for the attributes or scores.
[00204] Example
[00205] The example is based on a cube of seismic data with a size of 1426 by 1057 for 131 samples (also called voxels). Three geological attributes are computed: regularity, convergence and attachment. Regularity is computed with a relatively large window size of 61 by 41 by 41 samples. High regularity will demarcate planar strata, while low regularity will demarcate complex strata with complicated structure or complicated stratigraphy with erosion, entrenchment and filling. Given the large window used for computing regularity, intermediate values of regularity correspond to transitions between flat and complex strata, that is, areas where flat and complex strata are juxtaposed. Therefore, preferably, the association of flat strata with sealing rocks and complex strata with reservoirs leads to the identification of areas with intermediate regularity, such as those with juxtaposition of potential sealing rock reservoir.
[00206] The second geological attribute used in this example is convergence. No thinning in an area indicates that the strata are locally flat. High magnitudes of convergence or thinning indicate that the strata are locally changing their thickness and that the seismic surfaces are converging or diverging. The convergence is oriented and, therefore, a quantity of vector, however the present example discards orientation or signal and only uses the magnitude of convergence. Areas with marked convergence potentially contain stratigraphic traps, for example, by strata that taper (for example, Fig. 13) or are chamfered by a non-conformity (for example, Fig. 14). Thus, the convergence attribute differentiates the potential areas of coexistence of trap-reservoir.
[00207] The third geological attribute used in this example is an inclusion. Each seismic volume voxel is used as a seed point for the formation of a seismic surface that follows the seismic structure. The surface is examined to determine whether the seed point is located in a closed contour surrounding an elevated location on the surface, or in a closed contour surrounding a local low point or an open contour intersecting the volume limit. In the first case, the point is closed, meaning a floating fluid particle moving upwards from this location potentially being trapped in a tape. In the latter case, the floating particle can rise along the surface and leave the volume unchecked. As a result, the inclusion attribute outlines potential trap areas.
[00208] The three attributes used are not independent of each other. Intermediate regularity values refer to the sealing rock and reservoir; high convergence refers to reservoir and trap; and inclusion only refers to trape. Instead of demixing them by forming separate scores for sealing rock, reservoir and trap, attributes are thresholded to form binary scores for the potential presence of sealing rock reservoir, trap-reservoir and trap. Preferably, multiple attributes would be broken down into separate scores for the different elements of the hydrocarbon system and / or trap configuration, for example, by analysis of principle or grouping components. However, given the usual application in conjunction with the all-or-nothing logic provided by the door, the results of this preferable embodiment of the invention will be very similar in a much higher computational effort.
[00209] The three conditions defined will be met in different locations. To suppress small and often isolated areas, the labeling of the connected component is carried out through the intersection of the three scores to find larger, contiguous regions where samples of all three conditions are met simultaneously. Figure 20 shows a slice extracted from the three dimensional data moving through the system. For illustrative purposes, the system is shown in a cascade or serial mode, where the score of each attribute is used to sequentially destroy the voxels without having a particular aspect. Alternatively, the system could be implemented in a parallel mode by simultaneous intersection of all scores immediately. 200 represents a data slice of 131 slices. The first step is computation of the tissue or structure that is performed by skeletonizing the seismic horizon, followed by construction of a depth mapping volume, 201, where zebra lists are used to illustrate the detected tissue. The regularity attribute, 202, is computed and areas with low and high regularities are suppressed (light gray), while areas with intermediate regularity are emphasized (dark gray). These now-binary regularity scores are used to suppress seismic voxels with low or high regularity 203. The seismic tissue, 201, is used to compute the convergence attribute 204 which is converted into a threshold score. By suppressing seismic voxels with low convergence scores, more areas of seismic data 205 are destroyed. Seismic tissue 201 is still used to compute inclusion attribute 206, which is then used directly as a score. By eliminating seismic voxels without inclusion, further areas of seismic data 207 are reduced. Many voxels that obtain all three criteria are isolated or connected to very few similar ones. Thus, the final step is to consider the size of the remaining connected areas and suppress all but the largest 208. In this example, only four regions remain, one of which is a known hydrocarbon reservoir.
[00210] The foregoing patent application is directed to particular embodiments of the present invention for the purpose of illustrating it. It will be evident, however, to a person skilled in the art, that many modifications and variations in the embodiments described here are possible. All such modifications and variations are intended to fall within the scope of the present invention, as defined in the appended claims. Those skilled in the art will readily recognize that in practical applications of the invention, at least some of the steps of the present inventive method are performed on or with the aid of a computer, that is, the invention is implemented by computer.

权利要求:
Claims (17)
[0001]
1. Computer-implemented method to analyze a volume composed of data voxels (31) representing a subsurface region as to the presence of a particular hydrocarbon system or play (a conceptual model of a hydrocarbon accumulation style), comprising: dividing (3) the volume of seismic data to form a plurality of segments; and classify the plurality of segments as to the presence of a hydrocarbon system or the particular play based at least partially on the prospective scores (4) for the seismic data voxels in each segment; where the prospectivity score (4) is based on the computation of at least two attributes (2, 32) that refer to different elements of a hydrocarbon system or the particular play; characterized by the fact that: planing is applied to at least two attributes (2, 32), thereby generating geological attributes (2), and planing is applied across distances of at least ten seismic data voxels; and one of at least two attributes (2, 32) is related to the existence of a hydrocarbon trap and the existence of a hydrocarbon trap is predicted by spatially correlating the geological attributes and comparing them with a catalog of configurations of hydrocarbon trap (8).
[0002]
2. Method according to claim 1, characterized in that the partition (3) comprises combining contiguous individual voxels based on a projection attribute or a prospective score computed for each voxel.
[0003]
3. Method according to claim 1, characterized by the fact that the elements are selected from a group consisting of reservoir, sealing, trap, source, load, coverage, maturation, migration, accumulation, and distribution.
[0004]
4. Method according to claim 1, characterized by the fact that at least two elements are selected for computing the prospective score (4) for each voxel, and for each element at least one seismic attribute (2, 32) is selected and then an element score will be calculated for each selected element based on at least one seismic attribute (2, 32) selected, and the prospective score will be formed by combining the element's scores.
[0005]
5. Method according to claim 1, characterized by the fact that it still comprises determining from the seismic data a layer structure, called a tissue (1), representative of the subsurface region, and in which the planing to generate the geological attributes (2 , 32) comprises integrating or averaging the selected attributes along the fabric (1) to avoid mixing other strata.
[0006]
6. Method according to claim 5, characterized by the fact that planing along the fabric (1) is carried out by anisotropic, non-linear diffusion filtering or by integral convolution of the line.
[0007]
Method according to claim 6, characterized in that the tissue (1) is determined by seismic skeletonization.
[0008]
8. Method according to claim 1, characterized in that the at least two attributes (2, 32) comprise at least one of a group consisting of Hessian attributes, isocontour attributes, convergence (204), and confluence.
[0009]
Method according to claim 1, characterized in that the at least two attributes (2, 32) comprise at least one of a group consisting of wrap (206) and spill point; regularity (202); phase residues; texture; main component analysis anomalies provided with windows; play attributes; terminations; termination densities; diffusion maps; and their combinations.
[0010]
10. Method according to claim 1, characterized by the fact that the segments do not overlap.
[0011]
11. Method according to claim 1, characterized by the fact that the segments are classified based on a single score for each segment generated, combining the prospective scores for each voxel of the segment.
[0012]
12. Method according to claim 1, characterized in that each voxel of the seismic data volume is a separate segment.
[0013]
13. Method according to claim 1, characterized by the fact that the attribute values computed for different attributes (2, 32) are normalized on a common scale.
[0014]
14. Method according to claim 1, characterized by the fact that it still comprises estimating a confidence value for at least one prospective score (4).
[0015]
15. Method according to claim 2, characterized by the fact that the projection attribute is computed using principal component analysis provided with windows or diffusion mapping applied to at least one seismic attribute (2, 32).
[0016]
16. Computer program product, comprising an average usable computer having a computer-readable program code embedded in it, said computer-readable program code adapted to be executed to implement a method for analyzing seismic data (31) representing a region subsurface as to the presence of a hydrocarbon system or a particular play, which comprises: dividing (3) the volume of seismic data to form a plurality of segments; and classify the plurality of segments as to the presence of a hydrocarbon system or the particular play based at least in part on the prospective scores (4) for the seismic data voxels in each segment; where the prospectivity score (4) is based on the computation of at least two attributes (2, 32) that refer to different elements of a hydrocarbon system or the particular play; characterized by the fact that planing is applied to at least two attributes (2, 32), thereby generating geological attributes (2), and planing is applied across distances of at least ten seismic data voxels; and one of at least two attributes (2, 32) is related to the existence of a hydrocarbon trap and the existence of a hydrocarbon trap is predicted by spatially correlating the geological attributes and comparing them with a catalog of configurations of hydrocarbon trap (8).
[0017]
17. Method for producing hydrocarbons, characterized by the fact that it comprises: conducting a seismic inspection of a subsurface region; obtaining a prospective analysis of the seismic data (31) of the inspection in which the analysis was conducted by a method as defined in claim 1, which is incorporated herein by reference; drilling a well in the subsurface region based at least in part on the prospective analysis, and producing hydrocarbons from the well.

类似技术:

---

公开号 | 公开日 | 专利标题

---

BR112012028653B1|2020-11-10|method for seismic analysis of hydrocarbon systems

---

US8213261B2|2012-07-03|Method for geophysical and geological interpretation of seismic volumes in the domains of depth, time, and age

---

CN104614766B|2017-10-13|Seismic horizon skeletonization

---

US6823266B2|2004-11-23|Method for performing object-based connectivity analysis in 3-D seismic data volumes

---

US9874648B2|2018-01-23|Reservoir connectivity analysis in a 3D earth model

---

US20050171700A1|2005-08-04|Device and system for calculating 3D seismic classification features and process for geoprospecting material seams

---

Zhao et al.2018|Seismic attribute selection for unsupervised seismic facies analysis using user-guided data-adaptive weights

---

US11054537B2|2021-07-06|Feature index-based feature detection

---

US20200041692A1|2020-02-06|Detecting Fluid Types Using Petrophysical Inversion

---

US20200183031A1|2020-06-11|Automated seismic interpretation-guided inversion

---

US10677948B2|2020-06-09|Context based bounded hydrocarbon formation identification

---

EP2948795B1|2017-08-09|Fluid migration pathway determination

---

US10481297B2|2019-11-19|Fluid migration pathway determination

---

Zhou2020|Data-Driven Modeling and Prediction for Reservoir Characterization and Simulation Using Seismic and Petrophysical Data Analyses

---

Igbokwe2011|Stratigraphic interpretation of well-log data of the athabasca oil sands of alberta canada through pattern recognition and artificial intelligence

---

WO2021226613A1|2021-11-11|Structured representations of subsurface features for hydrocarbon system and geological reasoning

---

Leal Freitez0|Machine learning and fractal attributes applied to automatic detection and classification of geologic planes in resistivity image logs

同族专利:

---

公开号 | 公开日

---

CN102918423B|2016-09-07|

---

CN102918423A|2013-02-06|

---

EP2577356A4|2017-08-16|

---

RU2573166C2|2016-01-20|

---

AU2011258764A1|2012-12-06|

---

RU2012157782A|2014-07-10|

---

MY162927A|2017-07-31|

---

US9194968B2|2015-11-24|

---

EP2577356A1|2013-04-10|

---

US20130064040A1|2013-03-14|

---

BR112012028653A2|2016-08-09|

---

EP2577356B1|2020-09-16|

---

AU2011258764B2|2014-10-23|

---

CA2799504A1|2011-12-01|

---

CA2799504C|2018-01-02|

---

WO2011149609A1|2011-12-01|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

---

US4916615A|1986-07-14|1990-04-10|Conoco Inc.|Method for stratigraphic correlation and reflection character analysis of setsmic signals|

---

US5047991A|1989-04-28|1991-09-10|Schlumberger Technology Corporation|Lithology identification using sonic data|

---

US4992995A|1989-10-24|1991-02-12|Amoco Corporation|Methods for attenuating noise in seismic data|

---

US5274714A|1990-06-04|1993-12-28|Neuristics, Inc.|Method and apparatus for determining and organizing feature vectors for neural network recognition|

---

US5265192A|1990-09-20|1993-11-23|Atlantic Richfield Company|Method for the automated editing of seismic traces using an adaptive network|

---

GB9206676D0|1992-03-27|1992-05-13|Geco As|Method of locating hydrocarbon reserves|

---

GB9214482D0|1992-07-08|1992-08-19|Armitage Kenneth|Sequence property interpretation & risk analysis link|

---

US5444619A|1993-09-27|1995-08-22|Schlumberger Technology Corporation|System and method of predicting reservoir properties|

---

US5416750A|1994-03-25|1995-05-16|Western Atlas International, Inc.|Bayesian sequential indicator simulation of lithology from seismic data|

---

GB2322704B|1994-07-07|1998-12-09|Geco As|Method of Processing seismic data|

---

US5586082A|1995-03-02|1996-12-17|The Trustees Of Columbia University In The City Of New York|Method for identifying subsurface fluid migration and drainage pathways in and among oil and gas reservoirs using 3-D and 4-D seismic imaging|

---

US5539704A|1995-06-23|1996-07-23|Western Atlas International, Inc.|Bayesian sequential Gaussian simulation of lithology with non-linear data|

---

FR2738920B1|1995-09-19|1997-11-14|Elf Aquitaine|METHOD FOR AUTOMATIC SEISMIC FACIAL RECOGNITION|

---

US5671136A|1995-12-11|1997-09-23|Willhoit, Jr.; Louis E.|Process for seismic imaging measurement and evaluation of three-dimensional subterranean common-impedance objects|

---

US6052650A|1997-02-27|2000-04-18|Schlumberger Technology Corporation|Enforcing consistency in geoscience models|

---

US6466923B1|1997-05-12|2002-10-15|Chroma Graphics, Inc.|Method and apparatus for biomathematical pattern recognition|

---

US5940778A|1997-07-31|1999-08-17|Bp Amoco Corporation|Method of seismic attribute generation and seismic exploration|

---

US6498989B1|1997-08-11|2002-12-24|Trans Seismic International, Inc.|Method for predicting dynamic parameters of fluids in a subterranean reservoir|

---

WO1999040529A1|1998-02-04|1999-08-12|Biodynamic Research Corp.|System and method for acquiring and quantifying vehicular damage information|

---

GB9904101D0|1998-06-09|1999-04-14|Geco As|Subsurface structure identification method|

---

GB9819910D0|1998-09-11|1998-11-04|Norske Stats Oljeselskap|Method of seismic signal processing|

---

US6236942B1|1998-09-15|2001-05-22|Scientific Prediction Incorporated|System and method for delineating spatially dependent objects, such as hydrocarbon accumulations from seismic data|

---

ES2223156T3|1999-06-03|2005-02-16|Jason Geosystems B.V.|METHOD OF DETERMINATION OF SPACE CHANGES IN THE STRATIGRAPHY OF THE SUBSUELO STRUCTURE, LITHOLOGY AND CONTENT OF FLUIDS AND REDUCTION OF THE SEISMIC NOISE.|

---

DE19943325C2|1999-09-10|2001-12-13|Trappe Henning|Process for processing seismic measurement data with a neural network|

---

US6295504B1|1999-10-25|2001-09-25|Halliburton Energy Services, Inc.|Multi-resolution graph-based clustering|

---

RU2148166C1|1999-10-27|2000-04-27|СЕРГЕЕВ Андрей Борисович|Method of hydrocarbon deposits development|

---

US6226596B1|1999-10-27|2001-05-01|Marathon Oil Company|Method for analyzing and classifying three dimensional seismic information|

---

US6574566B2|1999-12-27|2003-06-03|Conocophillips Company|Automated feature identification in data displays|

---

FR2808336B1|2000-04-26|2002-06-07|Elf Exploration Prod|METHOD OF CHRONO-STRATIGRAPHIC INTERPRETATION OF A SEISMIC SECTION OR BLOCK|

---

US6363327B1|2000-05-02|2002-03-26|Chroma Graphics, Inc.|Method and apparatus for extracting selected feature information and classifying heterogeneous regions of N-dimensional spatial data|

---

US6618678B1|2000-05-26|2003-09-09|Jason Geosystems B.V.|Method of joint analysis and interpretation of the subsurface from multiple seismic derived layer property data sets|

---

US6625541B1|2000-06-12|2003-09-23|Schlumberger Technology Corporation|Methods for downhole waveform tracking and sonic labeling|

---

AU2001271490B2|2000-06-30|2005-08-04|Exxonmobil Upstream Research Company|Method for imaging discontinuities in seismic data using dip-steering|

---

GC0000235A|2000-08-09|2006-03-29|Shell Int Research|Processing an image|

---

FR2813959B1|2000-09-11|2002-12-13|Inst Francais Du Petrole|METHOD FOR FACILITATING THE RECOGNITION OF OBJECTS, ESPECIALLY GEOLOGICAL, BY A TECHNIQUE OF DISCRIMINANT ANALYSIS|

---

US6438493B1|2000-09-29|2002-08-20|Exxonmobil Upstream Research Co.|Method for seismic facies interpretation using textural analysis and neural networks|

---

US6950786B1|2000-10-10|2005-09-27|Schlumberger Technology Corporation|Method and apparatus for generating a cross plot in attribute space from a plurality of attribute data sets and generating a class data set from the cross plot|

---

US7006085B1|2000-10-30|2006-02-28|Magic Earth, Inc.|System and method for analyzing and imaging three-dimensional volume data sets|

---

GB2386455B|2000-12-18|2005-05-04|Schlumberger Holdings|Seismic signal processing method and apparatus for generating correlation spectral volumes to determine geologic features|

---

US6845245B2|2000-12-22|2005-01-18|Denso Corporation|Access parameter adaptation and packet data resource management using detailed mobile status information|

---

US7203342B2|2001-03-07|2007-04-10|Schlumberger Technology Corporation|Image feature extraction|

---

US6473696B1|2001-03-13|2002-10-29|Conoco Inc.|Method and process for prediction of subsurface fluid and rock pressures in the earth|

---

US6751558B2|2001-03-13|2004-06-15|Conoco Inc.|Method and process for prediction of subsurface fluid and rock pressures in the earth|

---

WO2003003054A2|2001-06-29|2003-01-09|Exxonmobil Upstream Research Company|Method for analysing dip in seismic data volumes|

---

US6853922B2|2001-07-20|2005-02-08|Tracy Joseph Stark|System for information extraction from geologic time volumes|

---

DE10142785C2|2001-08-31|2003-07-03|Henning Trappe|Method for determining local similarity from 3D seismic measurement data|

---

US6587791B2|2001-10-11|2003-07-01|Pioneer Natural Resources Usa, Inc.|System and method for assigning exploration risk to seismic attributes|

---

JP2003151933A|2001-11-19|2003-05-23|Tokyo Seimitsu Co Ltd|Wafer-polishing apparatus|

---

US6957146B1|2001-12-24|2005-10-18|Rdsp I, L.P.|System for utilizing seismic data to estimate subsurface lithology|

---

US7016550B2|2002-04-19|2006-03-21|Lockheed Martin Corporation|Scene-based non-uniformity offset correction for staring arrays|

---

FR2841344B1|2002-06-19|2005-04-29|Tsurf|METHOD, DEVICE AND PROGRAM PRODUCT FOR SMOOTHING SUBSURFACE PROPERTY|

---

US7188092B2|2002-07-12|2007-03-06|Chroma Energy, Inc.|Pattern recognition template application applied to oil exploration and production|

---

US20060184488A1|2002-07-12|2006-08-17|Chroma Energy, Inc.|Method and system for trace aligned and trace non-aligned pattern statistical calculation in seismic analysis|

---

US7162463B1|2002-07-12|2007-01-09|Chroma Energy, Inc.|Pattern recognition template construction applied to oil exploration and production|

---

US7308139B2|2002-07-12|2007-12-11|Chroma Energy, Inc.|Method, system, and apparatus for color representation of seismic data and associated measurements|

---

US20050288863A1|2002-07-12|2005-12-29|Chroma Energy, Inc.|Method and system for utilizing string-length ratio in seismic analysis|

---

US7184991B1|2002-07-12|2007-02-27|Chroma Energy, Inc.|Pattern recognition applied to oil exploration and production|

---

US7295706B2|2002-07-12|2007-11-13|Chroma Group, Inc.|Pattern recognition applied to graphic imaging|

---

GB2394050B|2002-10-07|2005-11-23| Westerngeco Seismic Holdings Limited|Processing seismic data|

---

US7453767B1|2003-11-25|2008-11-18|Michael John Padgett|Method for deriving a 3D GRAZ seismic attribute file|

---

US7463552B1|2003-11-25|2008-12-09|Michael John Padgett|Method for deriving 3D output volumes using filters derived from flat spot direction vectors|

---

US6810332B2|2003-01-31|2004-10-26|Chevron U.S.A. Inc.|Method for computing complexity, confidence and technical maturity indices for reservoir evaluations|

---

GB2400664B|2003-04-10|2005-05-25|Schlumberger Holdings|Extrema classification|

---

US6804609B1|2003-04-14|2004-10-12|Conocophillips Company|Property prediction using residual stepwise regression|

---

US7113869B2|2003-05-22|2006-09-26|Schlumberger Technology Corporation|Method for prospect identification in asset evaluation|

---

US6970397B2|2003-07-09|2005-11-29|Gas Technology Institute|Determination of fluid properties of earth formations using stochastic inversion|

---

US7243029B2|2003-08-19|2007-07-10|Apex Spectral Technology, Inc.|Systems and methods of hydrocarbon detection using wavelet energy absorption analysis|

---

US20050171700A1|2004-01-30|2005-08-04|Chroma Energy, Inc.|Device and system for calculating 3D seismic classification features and process for geoprospecting material seams|

---

WO2005119276A2|2004-05-27|2005-12-15|Exxonmobil Upstream Research Company|Method for predicting lithology and porosity from seismic reflection data|

---

US8077938B2|2005-02-09|2011-12-13|Siemens Medical Solutions Usa, Inc.|Diffusion tensor surface visualization|

---

WO2006100674A2|2005-03-21|2006-09-28|Yeda Research And Development Co. Ltd.|Detecting irregularities|

---

FR2884636B1|2005-04-15|2007-07-06|Earth Resource Man Services Er|PR0CEDE OF HIERARCHICAL DETERMINATION OF COHERENT EVENTS IN AN IMAGE|

---

US7953675B2|2005-07-01|2011-05-31|University Of Southern California|Tensor voting in N dimensional spaces|

---

US8625885B2|2006-03-23|2014-01-07|Intelliscience Corporation|Methods and systems for data analysis and feature recognition|

---

US7328107B2|2006-04-28|2008-02-05|Kjt Enterprises, Inc.|Integrated earth formation evaluation method using controlled source electromagnetic survey data and seismic data|

---

FR2909185B1|2006-11-27|2009-01-09|Inst Francais Du Petrole|METHOD OF STRATIGRAPHIC INTERPRETATION OF SEISMIC IMAGES|

---

US7454292B2|2007-04-13|2008-11-18|Saudi Arabian Oil Company|Inverse-vector method for smoothing dips and azimuths|

---

EP2168057B1|2007-07-16|2019-09-18|ExxonMobil Upstream Research Company|Geologic features from curvelet based seismic attributes|

---

CA2695137A1|2007-08-28|2009-03-05|Shell Internationale Research Maatschappij B.V.|Monitoring a region of interest in a subsurface formation|

---

CN103278847A|2007-11-14|2013-09-04|特拉斯帕克地球科学有限责任公司|Seismic data processing|

---

WO2009082545A1|2007-12-21|2009-07-02|Exxonmobil Upstream Research Company|Detection of features in seismic images|

---

US8803878B2|2008-03-28|2014-08-12|Schlumberger Technology Corporation|Visualizing region growing in three dimensional voxel volumes|

---

FR2930350B1|2008-04-17|2011-07-15|Inst Francais Du Petrole|PROCESS FOR SEARCHING FOR HYDROCARBONS IN A GEOLOGICALLY COMPLEX BASIN USING BASIN MODELING|

---

US8688425B2|2008-05-06|2014-04-01|Exxonmobil Upstream Research Company|Transport property data calculated from derivative seismic rock property data for transport modeling|

---

AU2009244695B2|2008-05-09|2013-08-22|Exxonmobil Upstream Research Company|Method for geophysical and stratigraphic interpretation using waveform anomalies|

---

CN102066980B|2008-05-22|2015-02-25|埃克森美孚上游研究公司|Seismic horizon skeletonization|

---

US8577613B2|2008-07-01|2013-11-05|Schlumberger Technology Corporation|Effective hydrocarbon reservoir exploration decision making|

---

CA2737415C|2008-11-06|2017-03-28|Exxonmobil Upstream Research Company|System and method for planning a drilling operation|

---

MY159169A|2008-11-14|2016-12-30|Exxonmobil Upstream Res Co|Windowed statistical analysis for anomaly detection in geophysical datasets|

---

EP2387730A1|2009-01-13|2011-11-23|ExxonMobil Upstream Research Company|Methods and systems to volumetrically conceptualize hydrocarbon plays|

---

US20100257004A1|2009-04-01|2010-10-07|Chervon U.S.A. Inc.|Method and system for conducting geologic basin analysis|

---

CA2764681C|2009-07-06|2017-11-21|Exxonmobil Upstream Research Company|Method for seismic interpretation using seismic texture attributes|

---

US8358561B2|2010-04-13|2013-01-22|Spectraseis Ag|Bayesian DHI for seismic data|

---

US9134443B2|2010-05-14|2015-09-15|Schlumberger Technology Corporation|Segment identification and classification using horizon structure|

---

US8457940B2|2010-07-29|2013-06-04|Schlumberger Technology Corporation|Model-consistent structural restoration for geomechanical and petroleum systems modeling|

---

GB201022128D0|2010-12-31|2011-02-02|Foster Findlay Ass Ltd|Active contour segmentation|

---

US20120197613A1|2011-01-31|2012-08-02|Chevron U.S.A. Inc.|Exploitation of self-consistency and differences between volume images and interpreted spatial/volumetric context|

---

US8972195B2|2011-01-31|2015-03-03|Chevron U.S.A. Inc.|Extracting geologic information from multiple offset stacks and/or angle stacks|

---

US8838391B2|2011-01-31|2014-09-16|Chevron U.S.A. Inc.|Extracting geologic information from multiple offset stacks and/or angle stacks|

---

US9121968B2|2011-01-31|2015-09-01|Chevron U.S.A. Inc.|Extracting geologic information from multiple offset stacks and/or angle stacks|

---

WO2012146894A2|2011-04-26|2012-11-01|Kenneth Rayvenor Lusty Armitage|Oil & gas exploration and production|

---

US20120322037A1|2011-06-19|2012-12-20|Adrienne Raglin|Anomaly Detection Educational Process|

---

US8954303B2|2011-06-28|2015-02-10|Chevron U.S.A. Inc.|System and method for generating a geostatistical model of a geological volume of interest that is constrained by a process-based model of the geological volume of interest|

---

US8843353B2|2011-08-25|2014-09-23|Chevron U.S.A. Inc.|Hybrid deterministic-geostatistical earth model|

---

US9128203B2|2011-09-28|2015-09-08|Saudi Arabian Oil Company|Reservoir properties prediction with least square support vector machine|

---

US20140081613A1|2011-11-01|2014-03-20|Austin Geomodeling, Inc.|Method, system and computer readable medium for scenario mangement of dynamic, three-dimensional geological interpretation and modeling|

---

US9261615B2|2012-06-15|2016-02-16|Exxonmobil Upstream Research Company|Seismic anomaly detection using double-windowed statistical analysis|

---

US20140052378A1|2012-08-14|2014-02-20|Chevron U.S.A. Inc.|Methods and corresponding software module for quantifying risks or likelihoods of hydrocarbons being present in a geological basin or region|

---

US20140118345A1|2012-10-26|2014-05-01|Chevron U.S. A. Inc.|System and method for analysis of trap integrity|

---

CN107169526B|2012-11-09|2020-10-16|加州理工学院|Method for automatic feature analysis, comparison and anomaly detection|

---

EP3042342A4|2013-09-04|2017-08-23|Neural ID LLC|Pattern recognition system|WO2011077300A2|2009-12-23|2011-06-30|Schlumberger Technology Bv|Processing of geological data|

---

EP2601574A1|2010-08-05|2013-06-12|ExxonMobil Upstream Research Company|Obtaining data from an earth model using functional decriptors|

---

WO2012102784A1|2011-01-26|2012-08-02|Exxonmobil Upstream Research Company|Method of reservoir compartment analysis using topological structure in 3d earth model|

---

US9798027B2|2011-11-29|2017-10-24|Exxonmobil Upstream Research Company|Method for quantitative definition of direct hydrocarbon indicators|

---

GB2498712A|2012-01-18|2013-07-31|Geco Technology Bv|Identifying features of interest in displays of geological objects|

---

EP2815255B1|2012-02-13|2017-03-01|Exxonmobil Upstream Research Company|System and method for detection and classification of seismic terminations|

---

WO2013144880A1|2012-03-29|2013-10-03|Westerngeco Llc|Analysis of geological objects|

---

US9261615B2|2012-06-15|2016-02-16|Exxonmobil Upstream Research Company|Seismic anomaly detection using double-windowed statistical analysis|

---

US20140118345A1|2012-10-26|2014-05-01|Chevron U.S. A. Inc.|System and method for analysis of trap integrity|

---

US10422900B2|2012-11-02|2019-09-24|Exxonmobil Upstream Research Company|Analyzing seismic data|

---

MX352438B|2012-11-08|2017-11-24|Total Sa|Method and device to process a three dimensional seismic image.|

---

FR2998397B1|2012-11-20|2015-07-24|Total Sa|METHOD FOR DETERMINING A REPRESENTATION OF A HYDROCARBON RESERVOIR|

---

US9523781B2|2012-12-27|2016-12-20|Schlumberger Technology Corporation|Normalization seismic attribute|

---

GB2511133B|2013-02-26|2018-08-29|Foster Findlay Ass Ltd|Enhanced visualisation of geologic features in 3D seismic survey data|

---

CA2904139C|2013-03-14|2021-09-07|Schlumberger Canada Limited|Direct fluid indicators in multiple segment prospects|

---

US10048396B2|2013-03-14|2018-08-14|Exxonmobil Upstream Research Company|Method for region delineation and optimal rendering transform of seismic attributes|

---

US9952340B2|2013-03-15|2018-04-24|General Electric Company|Context based geo-seismic object identification|

---

WO2014165512A1|2013-04-02|2014-10-09|Schlumberger Canada Limited|Processing seismic attributes using mathematical morphology|

---

CA2918268C|2013-08-16|2021-06-01|Michael David Ewing|Generating representations of recognizable geological structures from a common point collection|

---

AU2013399054B2|2013-08-30|2017-08-31|Landmark Graphics Corporation|A geostatistical procedure for simulation of the 3D geometry of a natural fracture network conditioned by well bore observations|

---

US10663609B2|2013-09-30|2020-05-26|Saudi Arabian Oil Company|Combining multiple geophysical attributes using extended quantization|

---

US10816688B2|2014-08-15|2020-10-27|Equinor Energy As|Method and apparatus for measuring seismic data|

---

US10995592B2|2014-09-30|2021-05-04|Exxonmobil Upstream Research Company|Method and system for analyzing the uncertainty of subsurface model|

---

CN104375179B|2014-11-05|2017-02-15|中国石油天然气集团公司|Method for looking for oil and gas reservoir based on TRAP-3D software|

---

CA2912626C|2014-11-05|2018-02-13|China National Petroleum Corporation|3d trap evaluation method of searching for oil-gas reservoir|

---

US11163092B2|2014-12-18|2021-11-02|Exxonmobil Upstream Research Company|Scalable scheduling of parallel iterative seismic jobs|

---

NL2014286B1|2015-02-12|2016-10-13|Biodentify B V|Computer supported exploration and production of heterogeneous distributed hydrocarbon sources in subsurface formations based on microbial prospecting.|

---

WO2016171778A1|2015-04-24|2016-10-27|Exxonmobil Upstream Research Company|Seismic stratigraphic surface classification|

---

CN104847264B|2015-05-05|2017-03-08|中国海洋石油总公司|A kind of method and apparatus realizing geosteering|

---

US10605940B2|2015-06-24|2020-03-31|Exxonmobil Upstream Research Company|Method for selecting horizon surfaces|

---

AU2015414299A1|2015-11-09|2018-04-12|Landmark Graphics Corporation|Modelling complex geological sequences using geologic rules and paleographic maps|

---

US10147173B2|2015-12-31|2018-12-04|General Electric Company|System and method for seismic data interpretation|

---

CN105676285B|2016-01-23|2017-09-05|北京诺克斯达石油科技有限公司|A kind of earthquake length Cyclic Analysis method based on PCA|

---

US10782431B2|2016-02-09|2020-09-22|Saudi Arabian Oil Company|Smoothing seismic data|

---

US10036820B2|2016-03-04|2018-07-31|General Electric Company|Expert guided knowledge acquisition system for analyzing seismic data|

---

US10677948B2|2016-03-04|2020-06-09|General Electric Company|Context based bounded hydrocarbon formation identification|

---

WO2017160273A1|2016-03-14|2017-09-21|Schlumberger Technology Corporation|Dip estimation via a modified structure tensor|

---

EP3529639A1|2016-10-18|2019-08-28|Services Petroliers Schlumberger|Determining subsurface layers using machine learning|

---

US10901118B2|2016-10-28|2021-01-26|Exxonmobil Upstream Research Company|Method and system for enhancing meshes for a subsurface model|

---

CA3055385A1|2017-03-14|2018-09-20|Saudi Arabian Oil Company|Collaborative sensing and prediction of source rock properties|

---

US10914852B2|2017-03-16|2021-02-09|International Business Machines Corporation|Unsupervised identification of seismic horizons using swarms of cooperating agents|

---

WO2018183992A1|2017-03-31|2018-10-04|Schlumberger Technology Corporation|Smooth surface wrapping of features in an imaged volume|

---

EP3646065A4|2017-06-26|2021-03-24|Services Pétroliers Schlumberger|Automatic feature extraction from seismic cubes|

---

CN107390268B|2017-07-05|2019-06-11|中国石油天然气股份有限公司|The recognition methods of low permeable glutenite rock diagenetic trap and device|

---

CA3076221A1|2017-09-22|2019-03-28|Saudi Arabian Oil Company|Estimating geological dip based on seismic data|

---

CN108374657B|2018-02-02|2021-10-19|东北石油大学|Automatic well breakpoint identification method|

---

BR112020022078A2|2018-05-03|2021-02-02|Chevron U.S.A. Inc.|system and method for mapping hydrocarbon matrix rock using seismic attributes|

---

US10983235B2|2018-09-14|2021-04-20|Saudi Arabian Oil Company|Characterizing depositional features by geologic-based seismic classification|

---

US11073637B2|2018-10-04|2021-07-27|Saudi Arabian Oil Company|Data structure for fast invasion percolation modeling software|

---

US11002721B2|2019-02-20|2021-05-11|Saudi Arabian Oil Company|Automated organic petrology|

---

CN109961437B|2019-04-04|2021-06-25|江南大学|Method for detecting significant fabric defects based on machine teaching mode|

---

US10928533B2|2019-04-24|2021-02-23|Saudi Arabian Oil Company|Identifying potential hydrocarbon traps in a subterranean region using recursive anisotropic erosion of seismic data|

---

US11249220B2|2019-08-14|2022-02-15|Chevron U.S.A. Inc.|Correlation matrix for simultaneously correlating multiple wells|

---

CN111025384A|2019-11-26|2020-04-17|中国石油天然气股份有限公司|Reservoir stratum prediction method and device based on waveform classification intersection fusion|

---

US11187826B2|2019-12-06|2021-11-30|Chevron U.S.A. Inc.|Characterization of subsurface regions using moving-window based analysis of unsegmented continuous data|

---

US11010969B1|2019-12-06|2021-05-18|Chevron U.S.A. Inc.|Generation of subsurface representations using layer-space|

---

US10984590B1|2019-12-06|2021-04-20|Chevron U.S.A. Inc.|Generation of subsurface representations using layer-space|

---

US11263362B2|2020-01-16|2022-03-01|Chevron U.S.A. Inc.|Correlation of multiple wells using subsurface representation|

---

US20210342724A1|2020-05-01|2021-11-04|Landmark Graphics Corporation|Petroleum play analysis and display|

---

US11193370B1|2020-06-05|2021-12-07|Saudi Arabian Oil Company|Systems and methods for transient testing of hydrocarbon wells|

---

EP3929630A4|2020-06-22|2021-12-29|Total Se|Computer implemented method for correcting a reservoir model of a reservoir geological formation based on seismic images|

---

EP3929631A1|2020-06-22|2021-12-29|Total Se|Method and system for analyzing a reservoir grid of a reservoir geological formation based on 4d seismic images|

---

CN111856579A|2020-07-31|2020-10-30|中国海洋石油集团有限公司|Method for comprehensively identifying basin edge sediment sector in deep water area of oil-gas-containing basin in south China sea|

---

US11202488B1|2020-08-10|2021-12-21|Saudi Arabian Oil Company|Sun shade|

---

CN112230301B|2020-09-18|2022-02-15|西南石油大学|Method for dividing cause types of deepwater water channels|

法律状态:
2018-12-26| B06F| Objections, documents and/or translations needed after an examination request according art. 34 industrial property law|

---

2019-10-01| B06U| Preliminary requirement: requests with searches performed by other patent offices: suspension of the patent application procedure|

---

2020-05-26| B09A| Decision: intention to grant|

---

2020-11-10| B16A| Patent or certificate of addition of invention granted|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 22/04/2011, OBSERVADAS AS CONDICOES LEGAIS. |

优先权:

---

申请号 | 申请日 | 专利标题

---

US34953410P| true| 2010-05-28|2010-05-28|

---

US61/349,534|2010-05-28|

---

PCT/US2011/033519|WO2011149609A1|2010-05-28|2011-04-22|Method for seismic hydrocarbon system analysis|

---