• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    An example of a method for tomographic migration velocity analysis may include collecting seismographic traces from a subterranean formation and using an initial velocity model to generate common image clusters and volume. depth image based, at least in part, on seismographic traces. A structure tensor can be calculated using said volume of the depth image for automated estimation of structural inclination and azimuth. A semblance can be generated using said plurality of common image groupings and said structure tensor. Depth residuals of the image can be automatically selected from said semblance. A ray tracing calculation can be performed on said initial velocity models using said structure tensor. An updated speed model can be generated with a tomographic inversion calculation, wherein said tomographic inversion calculation uses said plurality of residuals of the depth image and said ray tracing calculation.
    公开号:FR3027688A1
    申请号:FR1558994
    申请日:2015-09-24
    公开日:2016-04-29
    发明作者:Shengwen Jin；Shiyong Xu；Fan Xia；Richard Ottolini；Yiqing Ren
    申请人:Landmark Graphics Corp；
    IPC主号:
    专利说明:

    [0001] TOMOGRAPHIC SPEED ANALYSIS BY STRESS STRESSOR OF CONSTRAINTS CROSS REFERENCE TO RELATED APPLICATIONS [1] This application claims the benefit of US Provisional Application No. 62 / 068,161, entitled "STRUCTURE TENSOR CONSTRAINED TOMOGRAPHIC VELOCITY ANALYSIS", filed October 24, 2014 which is incorporated herein by reference for all purposes. BACKGROUND [2] The present invention generally relates to seismic exploration and specifically to tomographic velocity analysis using a structural tensor as a constraint. 1003] Seismology is used in exploration, archaeological studies and engineering projects that require geological information. Exploration seismology provides data that, when used in combination with other available geophysical, geological and borehole data, can provide information on the structure and distribution of rock types and their contents. . Such information is of great help in the search for water, geothermal reservoirs and mineral deposits such as hydrocarbons and ores. Most oil companies rely on exploration seismology to select sites in which to drill oil exploration wells. [4] Traditional seismology uses artificially generated seismic waves to map structures beneath the surface. Seismic waves propagate from a source down into the earth and reflect on the boundaries between structures below the surface. Surface receivers detect and record reflected seismic waves for further analysis. While some large-scale structures can often be perceived from a direct examination of recorded signals, recorded signals are usually processed using a sub-surface velocity model to eliminate distortions and reveal finer details in the image from below the surface. The quality of the image below the surface may depend on the accuracy of the speed model below the surface. [5] Speed analysis may include extracting velocity information from seismic data. A method for speed analysis includes a depth-of-stack migration technique, which has become an attractive tool for speed analysis not only because of its sensitivity to the speed model but also because of its ability to generate residual errors in the post-migration domain. A popular approach to Migration Rate Analysis (MVA) is the analysis of the residual curvature on a common image point grouping, which is based on a residual spread to measure the speed error. The analysis of residual curvature in areas of complex structure is coupled with a migration-inversion problem that can be analyzed from a tomographic point of view. 1006] Existing MVA tomographic processing methods require the selection step, including (1) the selection of the horizon in the depth of image volume for the estimation of tilt and azimuth information local and (2) the selection of residual spread in aggregations of the deep-migrated common image for the measurement of residual depth information. Manual selection can be tedious and time-consuming, especially in iterative processing and interpretation techniques. BRIEF DESCRIPTION OF THE FIGURES [7] A more complete understanding of the embodiments of the present invention and the advantages thereof can be gained by reference to the following description taken in conjunction with the accompanying figures, in which the similar reference numbers indicate similar characteristics. [8] Figure 1 is a diagram illustrating a side view of an environment of an illustrative marine seismic survey, according to aspects of the present disclosure. [9] Figure 2 is a diagram illustrating a top view of an illustrative marine seismic survey environment, according to aspects of the present disclosure. Figure 3 is a diagram illustrating an illustrative mid-point diagram that is derived from flip-flop shots received by a given chain, according to aspects of the present disclosure. FIG. 4 is a diagram illustrating an illustrative seismic survey recording system according to aspects of the present disclosure. Figure 5 is a diagram illustrating an illustrative trace set according to aspects of the present disclosure. Figure 6 is a diagram illustrating an illustrative volume of 3D data, according to aspects of the present disclosure. Figure 7 is a diagram illustrating an illustrative geometry snapshot, according to aspects of the present disclosure. Figure 8 is a flowchart illustrating a tomographic MVA method, according to aspects of the present disclosure. Figure 9 is a diagram illustrating eigenvectors in a structure tensor, according to aspects of the present disclosure. Figure 10a-e are diagrams illustrating obtaining structure information from a migrated image, according to aspects of the present disclosure. Figure 11 are diagrams illustrating the result of automatically selected depth residuals, according to aspects of the present disclosure. Figures 12a-b are diagrams illustrating an example of an in-line tilting angle superimposed on an inverted-time migrated image and superimposed ray trajectories on a speed model, respectively, according to aspects of the present disclosure. Figures 13a-b are diagrams illustrating the tilt and azimuth from the structure tensor of a SEAM data set, according to aspects of the present disclosure. Figures 14a-b are diagrams illustrating a comparison of a ray density coverage in a depth slice on the surface from a SEAM data set, according to aspects of the present disclosure. Figures 15a-j are diagrams illustrating updated speed, image and cluster comparisons from an exemplary SEAM dataset, according to aspects of the present disclosure. Figure 16 is a diagram illustrating an illustrative imaging system, according to aspects of the present disclosure. While embodiments of this disclosure have been illustrated and described and are defined by reference to exemplary embodiments of disclosure, such references do not imply a limit on disclosure, and no limitation of disclosure. the kind must not be deduced. The subject matter of the disclosed invention is capable of substantial modifications, alterations and equivalents in form and function, as will be apparent to those skilled in the relevant field who benefit from this disclosure. The illustrated and described embodiments of this disclosure are only examples, and are not an exhaustive description of the disclosure.
    [0002] DETAILED DESCRIPTION [0025] Illustrative embodiments of the present invention are described in detail below.  For the sake of clarity, all the features of an actual embodiment are not described in this specification.  It will, of course, be appreciated that in the development of any real embodiment, that many concrete decisions need to be made in order to achieve the specific objectives of the developers, such as compliance with constraints related to system or monetary considerations, which will vary from one implementation to another.  In addition, it will be appreciated that such a development effort can be complex and time-consuming, but would nevertheless be a routine undertaking for those skilled trades who benefit from this disclosure.  To facilitate a better understanding of the present disclosure, the following examples of certain embodiments are given.  In no case may the following examples be construed as limiting, or defining, the scope of the invention.  Embodiments of the present disclosure may be applicable to horizontal, vertical, deviated or otherwise non-linear wellbores in any type of subterranean formation.  Embodiments may be applicable to injection wells as well as production wells, including hydrocarbon wells.  Embodiments may be implemented with a tool that is suitable for testing, retrieving, and sampling along sections of the formation.  Some or all aspects of the present disclosure may be implemented in an information processing system or a computer system, both of which may be used interchangeably herein.  Examples of information processing systems include server systems, computer terminals, handheld devices, tablets, smart phones, and the like.  For purposes of this disclosure, an information processing system may include an instrumentality or an aggregate of instrumentalities operating to compute, classify, process, transmit, receive, retrieve, originate, switch, store, display , manifest, detect, record, reproduce, process or use any form of information, intelligence or commercial, scientific, control, or other purposes.  For example, an information processing system may be a personal computer, a networked storage device, or any other suitable device and may vary in size, shape, performance, functionality, and price.  The information processing system may comprise a RAM, one or more processing resources such as a central processing unit (CPU) or hardware or software control logic, a ROM and / or other types of non-volatile memory.  Additional components of the information processing system may include one or more disks, one or more network ports for communication with external devices as well as various input and output (I / O) devices such as a keyboard, a mouse and a video screen.  The information processing system may also include one or more buses operating to transmit communications between the various hardware components.  For the purposes of this disclosure, a computer-readable medium may include any instrumentality or aggregate of instrumentalities that may retain data and / or instructions for a period of time.  The computer readable medium may include, for example, without limitation, a storage medium such as a direct access storage device (e.g. , a hard disk or a floppy disk), a sequential access storage device (e.g. , a cassette), a compact disc, CD-ROM, DVD, RAM, ROM, EEPROM memory or flash memory; as well as communication media such as cables, optical fibers, microwaves, radio waves and other electromagnetic and / or optical media; and / or any combination of the foregoing elements.  As it is described here, an automated grid-based, 3D (MVA) tomographic migration velocity analysis approach can use a structure tensor as a constraint.  Examples of approaches may not require the manual selection of geological horizons in depth image volumes and residual smears in common migrated image clusters.  The structure tensor can be useful for estimating local tilt and azimuth information, which can be used as constraints for calculating Fréchet derivatives during tomographic inversion.  Aspects of the present disclosure may be included in an illustrative context such as marine seismic survey, such as that illustrated in Figures 1-5, although this disclosure is not limited to marine surveys.  At sea, seismic survey vessels can deploy a flute behind the ship as shown in Fig.  1.  Each flute 110 may trail behind the ship 100 as the ship moves forward (in the direction of arrow 102), and each flute includes multiple equally spaced receivers 114.  Each flute 110 may further include a programmable switcher 118 and programmable depth controls that draw the flute at an operating offset distance from the ship's path (see FIG.  2) and to a desired operating depth (Fig.  1).  The flutes 110 may be several kilometers long, and are generally constructed in sections of 25 to 100 meters in length that include groups of 35 or more uniformly spaced receivers.  Each flute 110 may comprise electrical or fiber optic cables for interconnecting the receivers 114 and seismic equipment on a ship 100.  The data can be digitized near the receivers 114 and transmitted to the ship 110 through the wiring at a rate of 7 (or more) million bits of data per second.  As shown in FIG.  1, the seismic survey ship 100 can also fire one or more sources 112.  Source 112 may be a pulse source or a source of vibration.  Receptors 114 used in marine seismology are commonly referred to as hydrophones, and are generally constructed with a piezoelectric transducer.  Various types of hydrophones are available such as disk hydrophones or cylindrical hydrophones.  The sources 112 and the receivers 114 are generally deployed below the surface of the ocean 104.  Processing equipment, such as information processing systems, aboard the control vessels of the operation of the sources and receivers and records the acquired data.  The seismic surveys can provide imagery data below the surface of the ocean 104 to reveal sub-surface structures such as the structure 106, which is below the ocean floor 108.  Analysts use imaging methods to process the data and map the topography of the layers beneath the surface.  The seismic survey data reveal various other characteristics of the subsurface layers that can be used to determine the locations of the oil and / or gas reservoirs.  In order to image the structure 106 under the surface, the source 112 can emit seismic waves 116 which are reflected when there are changes in the acoustic impedance caused by a structure below the surface 106 ( and other reflectors below the surface).  The reflected waves are detected by a receivers scheme 114.  By recording (as a function of time) incoming seismic waves 116 traveling from a source 112 to the structure 106 below the surface 114, an image of the structure 106 below the surface can be obtained after appropriate processing of the data.  [0035] FIG.  2, illustrates a top view (not to scale) of the seismic survey ship 100 which draws a set of flutes 110 and two sources 112.  When the ship 100 is advancing, the sources 112 can be triggered alternately in a pattern called "flip-flop".  Programmable controllers are used to allow more or less equal spacing between flutes.  The receivers at a given position on the flutes are associated with a common terrain trace file number or a common channel 202.  [0036] FIG.  3, illustrates a view from above of the illustrative positions of the source and receiver for two shots.  For a first shot, a source is triggered at position 302, and the illustrated portion of the receiver array is at position 304 (shown in dashed line).  For a second shot, a source is triggered at position 306 and the illustrated portion of the receiver array is at position 308 (shown in a solid line).  Assuming, for the moment, that the reflective structures below the surface are horizontal, the seismic waves that reach each of the 12 receivers are reflected from a position below the equidistant point between the positions of the source and the receiver.  Thus, the first shot produces reflections from below the 12 mid-point 311 points (shown in dotted line with vertical hatching), while the second shot produces reflections from below the 12 points at mid-point 310 ( illustrated in solid line with a horizontal hatch).  As an example, the vector 312 illustrates a propagation of the seismic energy from a plate 302 to a mid-point 314, and a vector 316 of equal length illustrates the propagation of the seismic energy to a position of the receiver .  For the second shot 306, the vectors 318 and 320 illustrate a similar propagation trajectory.  It should be noted that the mid-point 314 is one of the mid-point points that is "bombarded" with multiple shots, thus making more signal energy available from these areas when information from snapshots are processed and associated.  Seismic surveys (for land and sea) are generally designed to provide an equally distributed grid of mid-point points with a relatively high average number of hits for each mid-point.  [0037] FIG.  4 shows an illustrative seismic survey recording system having receivers 114 coupled to a bus 402 for communicating digital signals to the data recording circuit 406 on a survey ship 100.  Position information sensors and other parameter sensors 404 are coupled to the data recording circuit 406 to enable the data recording circuit to store additional information useful for interpreting the recorded data.  Illustratively, such additional information may include information on network orientation and speed information.  A versatile digital data processing system 408, which may include an information processing system, is illustrated as being coupled to a data recording circuit 406, and is also illustrated as being coupled through a bus 402 to positioning devices 410 and seismic sources 112.  The processing system 408 configures the operation of the recording circuit 406, the positioning devices 410 and the seismic sources 112.  The recording circuit 406 can acquire the high speed data stream (s) from the receivers 114 on a non-volatile storage medium such as a magnetic or optical disk storage array.  The positioning devices 410 (including programmable controllers and depth controls) can control the position of the receivers 114 and the sources 112.  The seismic recording system FIG.  4 may include additional components that are not specifically illustrated here.  Eg. each flute 110 could have an independent bus 402 for coupling with the data logging circuit.  The processing system 408 may include a user interface having a graphics screen and keyboard or other user input acceptance method, and may also include a network interface for communicating seismic survey data stored in a user interface. central computer with powerful computing resources to process seismic survey data.  [0040] FIG.  5 shows illustrative seismic signals, which may be called traces, detected and sampled by receivers 114.  The signals indicate a certain measure of seismic wave energy as a function of time (e.g. , displacement, velocity, acceleration, pressure), and they are digitized at high resolution (e.g. , 24 bits) with a programmable sampling rate.  Such signals can be grouped in different ways, and when they are grouped, they are called a "grouping".  Eg. , a "common mid-point grouping" is a group of traces that have a mid-point in a defined region.  A "snapshot grouping" is a group of traces recorded during a single trigger of the seismic source.  A "multi-cliché grouping" is a collection of clichés, often including all traces recorded along a navigation path during a marine seismic survey.  [0041] Although it is possible to relate the different waveforms recorded in the format illustrated in FIG. 5, side by side with a graph revealing large-scale structures beneath the surface, such structures are distorted and do not illustrate the finer structures.  In some embodiments, the raw waveforms illustrated in FIG.  5, can be processed to create a volume of the image of the depth, c. -to-d. , a 3D network of data values such as those illustrated in FIG.  6.  Depth image volume represents some seismic attributes across various depths and spatial orientations within a surveyed region.  The 3D network comprises cells of uniform size, each cell having a data value representing the seismic attribute for that cell.  Various seismic attributes may be represented, and in some embodiments each cell has multiple data values to represent multiple seismic attributes.  Examples of suitable seismic attributes include reflectivity, acoustic impedance, acoustic velocity, and density.  The volumetric data format lends itself more voluntarily to computer analysis and visual performance, and for this reason, the volume of the depth image can be called a "3D image" of the surveyed region.  [0042] FIG.  7, shows how various parameters are related to the geometry of an illustrative 2D snapshot (the case in 3D is similar).  The seismic energy propagates along a ray 702 from a seismic source to a target interface 704 and is reflected to a receiver along a ray 706.  At the point of reflection (represented elsewhere by coordinates (x, y, z) and abbreviated here as a vector {right arrow above (x)}), the surface 704 has a normal vector {right arrow over (n)} at an angle α to the vertical.  The incoming ray 702 and the reflected ray 706 are at equal angles (but opposite) "open" with respect to the normal vector.  The seismic trace data initially collected during a survey can be acquired depending on the location of the image, the location of the receiver and time, c. -ad. , P (s, r, t).  Traditionally, a change of variable is made in order to place this data in the area of the mid-point / offset / time point, c. -to-d. , P (m, h, t), where the mid-point m ---- (s + r) 12 and offset h = ls-r112.  By observing that these data represent the wave fields observed at the surface (z = 0), the wave field equation is used to extrapolate the wave field below the surface, a process known as migration.  An example of a migration technique comprises the following equations: P (m, h, t, z = 0) P (m, h, w, z = 0) (1) 9 P (m, h, w; z = 0) -> P (m, h, w, z) (2) P (m, h, w, z) - 313 (m, ph, r, z) (3) P (rn, ph, r; z) -) P (m, ph, r = 0; z) (4) [0045] Equation (1) represents a Fourier transform of all the data in order to place the acquired data at the level of the area (z = 0) in the mid-point / shift / frequency range.  Equation (2) represents the migration of the dataset using a well known double square root (DSR) to extrapolate an ende field.  Equation (3) represents a Radon transform, which may also be referred to as an inclined stack operation, data in the midpoint dot-p-tau domain.  The offset radius parameters p and tau may represent the slope and intersection of the inclined lines used to stack the data.  As indicated by equation (4), the definition of tau equal to zero provides a set of common angle domain image groupings, which can be visualized as a set of images P (m, z), each image being derived from the seismic energy striking the reflector at a different angle.  The shifted radius parameter ph is related to the local inclination a and the open angle Es by the equation: ph = 2 * S (rn, z) * cos a * sin 0 (5) where S (m, z) represents the slowness (the inverse of the acoustic velocity V (m, z)) in the vicinity of the reflector.  The tomographic MVA can be used to determine and / or refine a velocity model based on the depth mismatches in the common image clusters.  In the post-migrated angle domain, the seismic data P (m, ph, z) represents the depth positions of multiple images of the reflector location.  Using a good velocity-depth model V (m, z) in the migration generates flat CA1 clusters in the ph-z domain (ie, the reflectors appear as events at a constant depth z , whatever ph).  Otherwise, the depth residuals are present on the CAI groupings, which means that the depth of the event varies with ph).  In order to adapt the depth residual domain ph-z to the tomographic MVA approach, they are converted into travel time perturbations At (ph), which reflect the residual spread of a specular radius path.  Having chosen a reference depth, the depth residuals Az from the reference depth at a reflector location can be determined using a computational similarity between common images calculated at different angles in the image groupings. common area of angle.  The conversion of depth residual to path time perturbation in the ph-z domain can be expressed as At (ph) = Az, i4S2cos2 a-ph2 (6) where S represents the local slowness over the disturbance reflector and a represents the local inclination angle of the reflector.  Equation (6) can calculate the disruption of travel time caused by the extra length of the path a ray must travel due to the deviation of the depth.  The dependence of travel time disturbance on the angle of inclination of the reflector is small for small inclinations but becomes important for larger inclinations.  It should be noted that if the incident angle θ is desired it can be obtained without ray tracing using equation (5).  Therefore, the travel time disturbances calculated from the CAI groupings are insensitive to the ray path errors, allowing the use of a faster ray tracing algorithm.  Existing MVA tomographic processing methods, including the MVA tomographic processing method, may require one or more steps in which the values are manually selected by an engineer or technician.  The step may include, e.g. , the selection of a horizon in the volume of the depth image for the estimation of local inclination and azimuth information and the selection of the residual spread in the groupings of the common image migrated for depth to measure residual depth information.  Manual selection can be tedious and time-consuming, especially in iterative processing and interpretation techniques in which values are chosen at each iteration.  According to aspects of the present disclosure, an automated MVA tomographic migration approach can use a structure tensor as a constraint so that manual selection of geological horizons in depth image volumes and residual smears in recruitments. of the migrated common image are not necessary.  The structure tensor, e.g. , may be useful for estimating local inclination and azimuth information, which can be used as constraints for calculating Fréchet derivatives during tomographic inversion.  Figure 8 is a flowchart illustrating a tomographic MVA method 800, according to aspects of the present disclosure.  The method can begin at step 805 in which the seismographic traces in the form of pre-stacked data are collected.  Seismographic traces can be collected, eg. , using a seismic survey system similar to or different from that described with reference to FIG.  1.  The collection of seismographic traces may also include the reception of information at the level of a processing system or a processor of an information processing system, the previously collected seismographic traces coming from a medium on which traces have been previously stored.  This may include, e.g. , a memory device coupled to a processor, or to a server in a central data center.  Previously recorded tracks can be received, eg. , through one or more communication channels cabled or not.  In step 810, a deep migration can be performed based, at least in part, on the speed model 860.  Migration can, eg. , take the form of the example of the migration technique described above, but this migration technique is not intended to be limited, and may include other migration techniques that would be understood by one skilled in the art. the light of this disclosure.  In the first iteration of step 810, the speed model 860 may comprise an initial velocity model.  In subsequent iterations of step 810, the speed model 860 may comprise an updated speed model from step 855 (shown below).  The deep migration at step 810 may be used to determine a depth image volume 815 and common image groupings 830.  The depth image volume 815 may be the result of the depth migration of step 810 and may, but not necessarily, take a shape similar to the volume of the depth image described above with reference to the Fig.  6.  The common image groupings 830 may, but need not, include common angle domain image groupings determined using the method described above.  Other types of common image groupings are possible, as will be understood by a person skilled in the art in light of this disclosure.  In step 820, a structure tensor can be calculated from a depth image volume 815 to estimate structural tilt and azimuth information of the depth image volume. 815.  In some embodiments, the calculation of the structure tensor may, but not necessarily, compute smoothed Gaussian derivatives throughout the volume of the depth image 815, and then the proper decomposition can be determined using the following equation s =. 1. uuT +. ## EQU1 ## where ## EQU1 ##  Figure 9 is a diagram illustrating eigenvectors in a structure tensor, according to aspects of the present disclosure.  As shown in FIG.  9, the eigenvectors can define a coordinate system binomial tangential to the image gradient.  The calculation of the structure tensor can yield eigenvector volumes (3 Cartesian components each) and eigenvalues.  The tilt attribute volumes can be calculated from the normal of the tangent U; the magnitude of the inclination can be a Euclidean sum of 2 lateral Cartesian components; and the azimuth may be the arc tangent of these components.  Returning to FIG.  8, at step 835, the semblance can be calculated for the common image groupings 830, with a structure tensor of step 120 and the calculated inclination attributes used as a constraint.  In some embodiments, the structure tensor components may also be used as low signal masking or conflict tilting regions.  Although a number of structure-oriented semblances and planarity attributes may be used for eigenvalue ratios, in practice a threshold mask 4 may be adequate.  To illustrate it, Figs.  10a-e show a method for obtaining structure information from a migrated image, according to aspects of the present disclosure.  The tilt extraction can be calculated for a simple synthetic syncline 1000.  As illustrated, FIG.  10a may comprise a section through the middle of a depth image of synclinal 1000.  The structural tensors can be constructed from a migrated inverse time image of FIG.  10a.  Fig.  10b illustrates the attribute, the eigenvalue in the tangent-normal direction, which may be one of the components of the amplitude- and coherence-related structure tensor.  Fig.  10c illustrates a threshold mask profile 4, which can be created using FIG.  10b.  The mask profile of FIG.  10c can be created to eliminate low amplitude, non-coherent noise.  Fig.  10d illustrates a Cartesian component line of Uenhgne, and FIG.  10th illustrates the masked Uenligne.  Figs.  10d and 10e respectively illustrate line inclination angles calculated from the structure voltages before and after application of the mask profiles of FIG.  10c.  Returning to FIG.  8, at step 840, the depth residuals can be automatically selected on the residual similarities from step 835.  In some embodiments, an automatic selection algorithm may be used that selects functions for maximizing summation across the appearance values in multiple directions simultaneously based on an input guide function and a positive search range and negative with respect to the entry guide.  For residual smears after the migration, the entry guide may be a zero spread.  The selected spread can be constrained to be smooth in vertical directions and all spatial directions to avoid noisy selection.  Automatic simultaneous selection of all points together through global optimization can also avoid wild selections.  Figure 11 illustrates the result of automatically selected depth residuals according to aspects of the present disclosure.  As shown in FIG.  11, the selection result may be a "hyperplane" of speed (among which, FIG.  11 illustrates the projection of a single line).  In this way, the calculation can be completely automated and eliminates the need for manual selection of inclination and horizon at any stage.  A control can be exercised on this calculation by adjusting parameters such as the smoothing in the calculation of the tensor and the type of masking used.  In step 825, ray tracing and / or kernel sensitivity calculations may be performed using the structure tensor of step 820 as a constraint.  In some embodiments, the ray paths can be calculated using a dynamic ray tracing algorithm.  Given the angles of local reflection and azimuth, a pair of incident / reflected rays can be taken starting at the point of reflection.  When both rays reach the surface, the locations of the source and receiver can be determined, and the source-receiver shift and the surface-grabbing azimuth can be obtained.  Such ray tracing allows direct processing of migrated angle clusters.  In order to process migrated offset clusters, for each reflection point, the appropriate reflection angle and the local azimuth angle can be calculated to correspond to the expected surface shift and the take-up azimuth until all non-matches are minimized within the given tolerances.  According to the aspects of the present disclosure, FIG.  12a illustrates an example of an in-line tilt angle superimposed on an inverted-time migrated image, and FIG.  12b illustrates examples of superimposed ray paths on a velocity model.  [00591 Returning to FIG.  8, at step 845, a tomographic reversal can be performed with the ray tracing of step 825 and the automatically selected depth residuals of step 840.  In some embodiments, the inversion can be performed using the following equation: v (r) L (h) 2 cos a [cos y / L (h = 0)] - As = Az (h) ( 8) [L (h) 1 (11 = 0)] / cosi where, is a term reflecting the ray tracing of step 825, As is a term reflecting the disruption of the actual slowness, and AZ) is a term reflecting the automatically selected deep residuals of step 840.  In this embodiment, the calculated rays or the sensitivity nuclei can be stored on a sparse jacobierme matrix and the inversion system can be resolved using the conjugate-gradient method.  In this way, an updated speed model can be obtained.  If the speed model generated in step 845 represents the fmal speed model, it may be the output of the system at step 865.  On the other hand, if additional MVA iterations are desired to further update the velocity model, the velocity model 860 generated in step 845 can be provided at step 855 to be used as the new velocity model in the iterations. future (starting at step 810 with deep migration).  The advantages of the tomographic MVA method described in the present disclosure can be found with reference to its application to a data set of the SEG Advanced Modeling Corporation (SEAM).  First, an initial stacked image can be obtained by deep migration using an initial velocity model.  Then, with the calculation of the structural tensors of the initial stacked image, the inclination and the azimuth can be obtained.  In the ray tracing procedure, tilt and azimuth information, which associates stresses from the structure tensor, can be entered.  Depending on the result of the self-selected depth residual and the precise ray tracing, the updated velocity can be determined.  Figs.  13a-b illustrate the tilt and azimuth from the structure tensor of the example of the SEAM data set, according to aspects of the present disclosure.  Figs.  13a and 13b respectively show the tilt and azimuth results from the volume structure tensor of the initial image.  As shown in Figs.  13a-b, the results from the structure tensor may be more accurate in comparison with the ray tracing result with a tilt-azimuth assumption of zero.  Figs.  14a-b illustrate a comparison of radius density coverage in a surface depth slice from a SEAM data set, according to aspects of the present disclosure.  Specifically, FIG.  14a illustrates the radius density coverage in a depth slice on the surface with the stress of the structure tensor, and FIG.  14b illustrates the radius density coverage in the absence of the structure tensor constraint.  The improved radius coverage with the use of the structure tensor stress can be observed in FIG.  14a in comparison with FIG.  14b.  Figures 15a-j illustrate updated speed, image and cluster comparisons from an exemplary SEAM dataset, according to aspects of the present disclosure.  Specifically, FIG.  15a illustrates an initial speed; FIG.  15b illustrates an updated speed after a first iteration of the tomographic MVA in the absence of the structure tensor stress; FIG.  15c illustrates a refreshed speed after a first iteration of the tomographic MVA in the presence of the structure tensor stress; and FIG.  15d illustrates the true speed of the sample SEAM dataset.  As illustrated, the velocity updated with the stress of the structure tensor (Fig.  15c) is closer to true speed (Fig.  15d) after the first iteration compared to the unconstrained updated velocity (Fig.  15b).  The improved accuracy is also evident in the comparison among the stacked image (Fig.  15-h).  Specifically, FIG.  15th illustrates an image of the initial speed by default; FIG.  15f shows an image using a refreshed velocity after a first iteration in the absence of the structure tensor constraint; FIG.  15g shows an image using a refreshed velocity after a first iteration in the presence of the constraint of the structure tensor; and FIG.  15h shows an image of the true speed of the sample SEAM dataset.  As illustrated, the image of the updated velocity with the stress of the structure tensor (Fig.  15g) is closer to true speed (Fig.  15h) after the first iteration compared with the velocity updated without constraint (Fig.  15f).  The improvement of the accuracy is all the more evident in two examples of common image grouping data sets (FIG.  15i and FIG.  15j).  Fig.  15i illustrates 4 offset image grouping boards 16 at location A in FIG.  15th.  From left to right, the 4 boards in Fig.  15i illustrate a grouping of staggered images with a default initial velocity, a discounted velocity in the absence of a structure tensor constraint, a velocity updated in the presence of a structure tensor constraint, and the true velocity of the dataset.  Fig.  15j illustrates 4 similar boards for shifted image groupings at a location B in FIG.  15th.  As illustrated, in the two examples of image grouping datasets, the updated velocity with a structure tensor constraint is closer to the true velocity than the updated velocity in the absence of a tensor stress. structure.  Some or all of the steps of the illustrative method described above with respect to FIG.  8 may comprise software steps performed in an information processing system.  Software may be characterized by a set of instructions stored on a computer-readable medium that, when executed by a processor, allows the processor to perform certain functions.  Fig.  16 shows an illustrative computer system 900 in which the illustrative method can be realized.  As illustrated, a personal workstation 902 is coupled through a Local Area Network (LAN) 904 to one or more multiprocessor computers 906, which are, in turn, coupled via the LAN to one or more shared storage units 908. .  Workstation 902 and computers 906 may include information processing systems.  The personal workstation 902 provides a user interface for the processing system, allowing a user to download data from the server to the system, retrieve and view image data from the system, and configure and monitor the operation of a treatment system.  The personal workstation 902 may be in the form of a desktop computer with a graphical display that graphically displays the survey data and 3D images of the surveyed region, and with a keyboard that allows the user to move files and run processing software.  The LAN 904 allows high speed communication between 906 multiprocessor computers and 902 personal workstations.  The LAN 904 can take the form of an Ethernet network.  Multiprocessor computer (s) 906 provide parallel processing capability for relatively fast conversion of seismic trace signals in a probed region image.  Each computer 906 includes multiple processors 912, a distributed memory 914, an internal bus 916, is a LAN interface 920.  Each processor 912 operates on an allocated portion of the input data to produce a partial image of the surveyed seismic region.  A distributed memory module 914 is associated with each processor 912 which stores the conversion software and a work data set for use by the processor.  An internal bus 916 allows inter-processor communication and communication with the LANs through the interface 920.  The communication between the processors in the different computers 906 can be provided by the LAN 904.  The shared storage units 908 may be large standalone information storage units that utilize a magnetic disk medium for nonvolatile data storage.  In order to improve access speed and reliability, the shared storage units 908 can be configured as a redundant disk array.  The shared storage units 908 store an initial volume of velocity data and clusters of snapshots from a seismic survey.  In response to a request from a workstation 902, the image volume data can be retrieved by computers 906 and fed to a workstation for conversion to a graphics image that will be displayed to a user.  An exemplary method for speed analysis by tomographic migration may include collecting seismographic traces from a subterranean formation and using an initial velocity model to generate common image groupings and a depth image volume based, at least in part, on the seismographic traces.  A structure tensor can be calculated with the depth image volume for automated estimation of structural tilt and azimuth.  A semblance can be generated using said plurality of common image groupings and said structure tensor.  Depth residuals of the image can be automatically selected from said semblance.  A ray tracing calculation can be performed on said velocity models using said structure tensor.  An updated speed model can be generated with a tomographic inversion calculation, wherein said tomographic inversion calculation uses said plurality of residuals of the depth image and said ray tracing calculation.  In some embodiments described in the preceding paragraph, the collection of seismographic traces comprises the emission of at least one seismic wave, and the reception of a reflection of the at least one seismic wave.  In some embodiments described in this clause, the use of an initial velocity model to generate the plurality of common image clustering and depth image volume based, at least in part, on Seismographic traces include performing a deep migration on the seismographic traces.  In some embodiments described in the preceding paragraph, the generation of the semblance using said plurality of common image groupings and said structure tensor comprises generating a semblance using the structure tensor as a constraint.  In some embodiments described in the preceding paragraph, the automatic selection of a plurality of image depth residuals from said semblance includes automatically selecting a plurality of image depth residuals using a selection algorithm. automatic that maximizes the appearance values in multiple directions based on an input guide function and a positive and negative search range relative to the input guide.  In certain embodiments described in the preceding paragraph, the calculation of the structure tensor using said depth image volume for the automated estimation of the structural inclination and the azimuth comprises the computation of smoothed Gaussian derivatives. in the volume of the depth image.  In some embodiments described in the preceding 3 paragraphs, the method may also include using the updated speed model to generate an updated plurality of common image clustering and a depth image volume.  In certain embodiments described in the preceding paragraph, the method may also comprise the calculation of an updated structure tensor using said updated depth image volume for the updated automated estimation of the structural inclination and of the azimuth; generating an updated semblance using said plurality of common image clustering and the updated structure tensor; automatically selecting a plurality of depth image residuals from said plurality of similarities; performing a ray tracing calculation on said updated velocity model using said updated structure tensor; and the generation of a second updated velocity model with computation of the tomographic inversion.  In some embodiments described in the preceding paragraphs, the method may also include determining one or more features of the training based, at least in part, on the updated speed model.  In some embodiments described in the preceding paragraph, one or more features of the formation include strata boundaries of the formation.  An example system may include a seismic survey system with at least one seismic source and at least one seismic sensor, and an information processing system comprising a processor and a memory device coupled to the processor.  The memory device may include a set of instructions which, when executed by a processor, allows the processor to collect seismographic traces from a subterranean formation, and use of an initial velocity model to generate a plurality of common image groupings and an image depth volume based, at least in part, on the seismographic traces.  The instruction set may also enable the processor to compute a structure tensor using said depth image volume for automated estimation of the structural tilt and azimuth, and to generate a semblance using said plurality of clusters common image and said structure tensor.  The instruction set may also allow the processor to automatically select a plurality of depth residuals of the image from said semblance; performing a ray tracing calculation on said initial velocity models using said structure tensor; and generating an updated speed model with a tomographic inversion calculation, wherein said tomographic inversion calculation uses said plurality of image depth residuals and said ray tracing calculation.  In some embodiments described in the preceding paragraph, the seismographic traces comprise at least one seismic wave received at the level of the at least one seismic sensor, in which the at least one seismic wave has been generated by at least one a seismic source and reflected by an underground formation.  In some embodiments described in this clause, the instruction set that allows the processor to use the initial velocity model to generate the plurality of common image clusters and the volume volume of the depth-based image, at least in part, on the seismographic traces also allows the processor to perform a deep migration on the seismographic traces.  In some embodiments described in the preceding paragraph, the set of instructions that allows the processor to calculate the structure tensor using said depth image volume the automated estimation of the structural inclination and the azimuth allows the processor to compute Gaussian derivatives smoothed in the volume of the depth image.  In some embodiments described in the preceding 3 paragraphs, the set of instructions that allows the processor to generate the semblance using the plurality of common image clustering and said structure tensor also allows the processor to generate the semblance using the structure tensor as a constraint.  In some embodiments described in the preceding 3 paragraphs, the instruction set which allows the processor to automatically select a plurality of image depth residuals from said semblance also enables the processor to automatically select a plurality of residuals. image depth utilizing an automatic selection algorithm that maximizes the appearance values in multiple directions, based on an input guide function and a positive and negative search range for the input guide.  In some embodiments described in the previous 3 paragraphs, the instruction set also allows the processor to use a refreshed velocity model to generate an updated plurality of common image clustering and a depth image volume.  In certain embodiments described in the preceding paragraph, the instruction set also allows the processor to calculate an updated structure tensor using said updated depth image volume for the updated automated estimation of the structural inclination. and azimuth; generating an updated semblance using said plurality of common image groupings and said updated structure tensor; automatically selecting a plurality of updated depth image residuals from said plurality of similarities; performing a ray tracing calculation on said updated velocity model using said updated structure tensor; and the generation of a second updated velocity model with computation of the tomographic inversion.  In some embodiments described in the preceding paragraphs, the instruction set also allows the processor to determine one or more characteristics of the training based, at least in part, on the updated speed model.  In some embodiments described in the preceding paragraph, one or more characteristics of the formation comprise strata boundaries of the formation.  Therefore, the present invention is well suited to achieve the purposes and obtain the advantages mentioned herein as well as those inherent in the present disclosure.  The particular embodiments disclosed above are illustrative only, since the present invention may be modified and practiced in a different but equivalent manner which will be apparent to those skilled in the art who benefit from the teachings of the present disclosure.  Furthermore, no limitation is contemplated to the construction or design details described herein, other than those described in the appended claims.  It is thus obvious that the particular illustrative embodiments described above may be altered or modified and any such variations are considered to be within the scope or spirit of the present invention.  But also, the terms in the claims have their plain and ordinary meaning, except in the case of explicit mention and clear definition of the plaintiff.  The undefined articles "a" or "an" are used in the claims and are both defined herein to mean one or more elements that they introduce.
    权利要求:
    Claims (10)
    [0001]
    REVENDICATIONS1. A method of analyzing tomographic migration velocity, comprising: collecting seismographic traces from a subterranean formation; using an initial velocity model to generate the plurality of common image clusters and a depth image volume based, at least in part, on the seismographic traces; calculating a structure tensor using said volume of the depth image for automated estimation of the structural inclination and the azimuth; generating a semblance using said plurality of common image groupings and said structure tensor; automatically selecting a plurality of residuals from the depth of the image from said semblance; performing a ray tracing calculation on said initial velocity models using said structure tensor; and generating an updated speed model with a tomographic inversion calculation, wherein said tomographic inversion calculation uses said plurality of depth image residuals and said ray tracing calculation.
    [0002]
    The method of claim 1, wherein: using the initial velocity model to generate the plurality of common image depth and the image volume of the depth based, at least in part, on the seismographic traces includes performing a deep migration on seismographic traces; and calculating the structure tensor using said depth image volume for automated estimation of the structural tilt and azimuth comprises computing smoothed Gaussian derivatives in the volume of the depth image.
    [0003]
    The method of claim 1, wherein: generating the semblance using said plurality of common image groupings and said structure tensor comprises generating the semblance using the structure tensor as a constraint; and automatically selecting a plurality of image depth residuals from said semblance includes automatically selecting a plurality of image depth residuals using an automatic selection algorithm that maximizes the appearance values in multiples. directions based on an input guide function and a positive and negative search range for the entry guide.
    [0004]
    The method of claim 1, further comprising: using the updated speed model to generate an updated plurality of common image clustering and a depth image volume; calculating an updated structure tensor using said depth image volume for the up-to-date automated estimation of the structural inclination and the azimuth; generating a present-day semblance using said plurality of common image groupings and said updated structure tensor; 24automatically selecting a plurality of image depth residuals updated from said updated semblance; performing a ray tracing calculation on said updated initial velocity model using said updated structure tensor; and the generation of a second updated velocity model with computation of the tomographic inversion.
    [0005]
    The method of claim 1, further comprising determining the stratum limits of the formation based, at least in part, on the updated velocity model.
    [0006]
    A system, comprising: a seismic survey system comprising at least one seismic source and at least one seismic sensor; an information processing system comprising a processor and a memory device coupled to the processor, the memory device containing a set of instructions which, when executed by the processor, enables the processor to perform the following steps: collection of seismographic traces from an underground formation; using an initial velocity model to generate a plurality of common image clusters and a depth image volume based, at least in part, on the seismographic traces; calculating a structure tensor using said volume of the depth image for automated estimation of structural inclination and azimuth; generating a semblance using said plurality of common image groupings and said structure tensor; 25for automatically selecting a plurality of residuals of the depth of the image from said semblance; performing a ray tracing calculation on said initial velocity models using the structure tensor; and generating an updated velocity model with a tomographic inversion calculation, wherein said tomographic inversion calculation uses said plurality of depth image residuals and said ray tracing calculation.
    [0007]
    The system of claim 6, wherein: the instruction set that allows the processor to use the initial velocity model to generate the plurality of common image clusters and the volume volume of the depth-based image, at least in part, on the seismographic traces also allows the processor to perform a deep migration on the seismographic traces; and the instruction set which allows the processor to calculate the structure tensor using said depth image volume for the automated estimation of the structural inclination and the azimuth comprises the calculation of Gaussian derivatives smoothed in the volume of the image of depth.
    [0008]
    The system of claim 6, wherein: the instruction set that allows the processor to generate the semblance using the plurality of common image clustering and said structure tensor also enables the processor to generate the semblance using the tensor structure as constraint; and the instruction set which allows the processor to automatically select a plurality of image depth residuals from said semblance also enables the processor to automatically select a plurality of image depth residuals using an automatic selection algorithm which maximizes look-up values in multiple directions, based on an input guide function and a positive and negative search range for the input guide.
    [0009]
    The system of claim 6, wherein the instruction set also allows the processor to use a refreshed velocity model to generate an updated plurality of common image clustering and a depth image volume; calculating an updated structure tensor using said volume of the updated depth image for automated estimation of the structural tilt and azimuth; generating updated similarity using said plurality of common image groupings and said updated structure tensor; automatically selecting a plurality of image depth residuals updated from said updated semblance; performing a ray tracing calculation on said updated initial velocity model using said updated structure tensor; and the generation of a second updated velocity model with computation of the tomographic inversion.
    [0010]
    The method of claim 6, wherein the instruction set also allows the processor to determine strata boundaries of the training based, at least in part, on the updated speed model. 27
    类似技术:
    公开号 | 公开日 | 专利标题
    FR3027688A1|2016-04-29|
    AU2009344283B2|2013-09-12|Seismic imaging systems and methods employing a fast target-oriented illumination calculation
    CA2670384C|2015-03-10|Joint inversion method for seismic data on different time scales
    US7952960B2|2011-05-31|Seismic imaging with natural Green's functions derived from VSP data
    FR3031210B1|2019-08-02|SEISMIC ELASTIC WAVE SIMULATION FOR TRANSVERSALLY INCLINED ISOTROPIC ENVIRONMENT USING DECADED LEBEDEV GRID
    CN108139499A|2018-06-08|The full wave field inversion of Q- compensation
    US9632192B2|2017-04-25|Method of processing seismic data by providing surface offset common image gathers
    WO2004034088A2|2004-04-22|Method and system for limited frequency seismic imaging
    AU2014249423B2|2017-07-06|Anisotropy analysis using direct and reflected arrivals in seismic survey data
    Biondi et al.2004|3D angle‐domain common‐image gathers for migration velocity analysis
    US10310117B2|2019-06-04|Efficient seismic attribute gather generation with data synthesis and expectation method
    US20170115411A1|2017-04-27|Joint inversion of seismic data
    US9952341B2|2018-04-24|Systems and methods for aligning a monitor seismic survey with a baseline seismic survey
    Hloušek et al.2015|Improved structural characterization of the Earth's crust at the German Continental Deep Drilling Site using advanced seismic imaging techniques
    US9658354B2|2017-05-23|Seismic imaging systems and methods employing correlation-based stacking
    Di et al.2018|3D structural-orientation vector guided autotracking for weak seismic reflections: A new tool for shale reservoir visualization and interpretation
    US20160266266A1|2016-09-15|Systems and methods employing upward beam propagation for target-oriented seismic imaging
    EP3570074A1|2019-11-20|Method for detecting geological objects in an image
    Thiel et al.2019|Comparison of acoustic and elastic full‐waveform inversion of 2D towed‐streamer data in the presence of salt
    FR3032532A1|2016-08-12|
    Nanda2021|Seismic modelling and inversion
    US20210374465A1|2021-12-02|Methodology for learning a similarity measure between geophysical objects
    US20200333489A1|2020-10-22|Unified continuous seismic reservoir monitoring
    US20220066059A1|2022-03-03|Method and system that uses an anisotropy parameter to generate high-resolution time-migrated image gathers for reservoir characterization, and interpretation
    US20210215842A1|2021-07-15|Identifying geologic features in a subterranean formation using seismic diffraction and refraction imaging
    同族专利:
    公开号 | 公开日
    WO2016064483A1|2016-04-28|
    AR102404A1|2017-02-22|
    CA2964893C|2019-04-02|
    NO20170626A1|2017-04-12|
    FR3027688B1|2018-07-13|
    GB2546046A|2017-07-05|
    US20160299243A1|2016-10-13|
    CA2964893A1|2016-04-28|
    US9869783B2|2018-01-16|
    AU2015337052A1|2017-05-04|
    GB201706131D0|2017-05-31|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US20020049540A1|2000-08-07|2002-04-25|Dimitri Bevc|Velocity analysis using angle-domain common image gathers|
    US20110075516A1|2009-09-25|2011-03-31|Halliburton Energy Services, Inc.|Seismic Imaging Systems and Methods Employing Tomographic Migration-Velocity Analysis Using Common Angle Image Gathers|
    US5157637A|1992-03-05|1992-10-20|Conoco Inc.|Method of structural traveltime tomography|
    FR2918178B1|2007-06-29|2009-10-09|Inst Francais Du Petrole|METHOD FOR ADJUSTING A SEISMIC WAVE SPEED MODEL BASED ON INFORMATION RECORDED AT WELLS|
    US8830788B2|2011-02-24|2014-09-09|Landmark Graphics Corporation|Sensitivity kernal-based migration velocity analysis in 3D anisotropic media|
    AU2014200118A1|2013-01-11|2014-07-31|Cgg Services Sa|Dip tomography for estimating depth velocity models by inverting pre-stack dip information present in migrated/un-migrated pre-/post-stack seismic data|US10217018B2|2015-09-15|2019-02-26|Mitsubishi Electric Research Laboratories, Inc.|System and method for processing images using online tensor robust principal component analysis|
    US10451765B2|2016-05-06|2019-10-22|Baker Hughes, A Ge Company, Llc|Post-well reservoir characterization using image-constrained inversion|
    CN109116413B|2018-07-30|2022-02-18|中国石油化工股份有限公司|Imaging domain stereo chromatography velocity inversion method|
    CN109188513B|2018-09-30|2020-03-10|中国石油天然气股份有限公司|Method and device for generating depth domain data volume and storage medium|
    RU2710972C1|2019-10-28|2020-01-14|Общество с ограниченной ответственностью «Сейсмотек»|Multivariate tomography method of seismic survey data|
    CN111239818B|2020-02-12|2022-02-25|成都理工大学|Ancient landform analysis method based on three-dimensional dip angle attribute body correction|
    法律状态:
    2016-07-25| PLFP| Fee payment|Year of fee payment: 2 |
    2017-07-26| PLFP| Fee payment|Year of fee payment: 3 |
    2017-11-03| PLSC| Search report ready|Effective date: 20171103 |
    2018-07-18| PLFP| Fee payment|Year of fee payment: 4 |
    2019-09-26| PLFP| Fee payment|Year of fee payment: 5 |
    2021-06-11| ST| Notification of lapse|Effective date: 20210506 |
    优先权:
    申请号 | 申请日 | 专利标题
    US201462068161P| true| 2014-10-24|2014-10-24|
    [返回顶部]