Methods, systems, and media for detecting hidden geological anomalies based on wavefield decomposition

Abstract

This invention provides a method, system, and medium for detecting concealed geological anomalies based on wavefield decomposition. The method includes: obtaining high signal-to-noise ratio raw data; randomizing the raw data according to the scale of concealed geological anomalies; performing wavefield decomposition on the randomized data using the Radon wavefield decomposition method or the frequency-spatial-domain wavefield decomposition method; obtaining layered wavefield (normal field) and lateral small-scale geological body wavefield (anomaly field) data, and outputting them. This solution solves the problem of fine description and characterization of the wavefields of irregular geological bodies such as fractures, fracture-dissolution bodies, and fracture-cavities. Furthermore, it can directly detect small-scale, irregular, concealed oil and gas reservoirs hidden in layered wavefields, providing an effective solution to the challenge of directly detecting these types of oil and gas reservoirs.

Description

Technical Field

[0001] This invention relates to the fields of wavefield data processing and oil and gas exploration technology, and in particular to a method, system and medium for detecting hidden geological anomalies based on wavefield decomposition. Background Technology

[0002] With the deepening development of oil and gas exploration and development in my country, current exploration work has revealed that oil and gas reservoirs not only possess layered characteristics, but also exhibit a wider variety of irregular features, including those affected by layered wave field interference, concealment, and submersion. These diverse types of oil and gas reservoirs include traditional layered reservoirs as well as various non-layered reservoirs.

[0003] Early explorations primarily uncovered layered oil and gas reservoirs, namely sandstone reservoirs within conventional marine and terrestrial strata. These reservoirs could be effectively detected using simple techniques based on layered wavefield theory. In recent years, non-layered oil and gas reservoirs have been discovered, exhibiting a diverse range of types, including: cave-fracture reservoirs in carbonate weathering crusts, fracture-dissolution reservoirs formed by caves and fractures controlled by carbonate faults, bioherms, oolitic limestone, fractured gas reservoirs, igneous rock reservoirs, dolomite fracture-fracture + dissolution fracture-pore network gas reservoirs, and reservoirs in channels, fans, and subsurface rivers.

[0004] The reservoir space of diversified oil and gas reservoirs is no longer limited to the characteristics of infinite lateral extension, finite vertical scale, and regular distribution. Instead, it is more characterized by finite lateral scale, varying vertical scale, and irregular and non-uniform distribution. This includes: ① two-dimensional bodies with one infinite lateral scale and another finite lateral and vertical scale; ② columnar bodies such as sinkholes and collapse columns with finite lateral scale and "infinite" vertical scale; ③ three-dimensional bodies with finite vertical and lateral scales, such as dissolution caves and scour caves; ④ irregular oil and gas-bearing sand bodies with finite vertical and lateral scales in layered sandstone reservoirs; ⑤ heterogeneous and irregular bodies composed of fracture networks such as fissures, cracks, and dissolution pores in homogeneous (quasi-homogeneous) media; and ⑥ various forms of irregular geological bodies such as rock masses, reefs, salt domes, gypsum salt, and rock salt.

[0005] Traditional seismic layered wavefield theory is based on seismic layered structural models. Seismic data acquisition, processing, interpretation, inversion, and hydrocarbon prediction are largely built upon these models. For example: ① Post-stack impedance inversion and pre-stack elastic inversion both rely on the variation of single-interface reflected wave amplitudes based on the layered structural model to derive the wave impedance and elastic parameters of the media above and below the interface; ② Time-domain and time-frequency-domain attribute analyses both focus on the amplitude and variation of reflected waves from a single thin layer (a thin layer with opposite reflection coefficient polarities at the top and bottom interfaces) within a homogeneous medium, based on the layered structural model. Practice shows that these techniques can produce biases when detecting thin interlayered structures, and they are even more deficient in the study of irregular wavefields. In fact, apart from the overt irregular wavefields generated by some irregular bodies that can be studied independently, the wavefields of other types of irregular bodies either interfere with and become blurred by the layered wavefield, or are masked by the layered wavefield, forming hidden hydrocarbon reservoirs. These are difficult to study independently and can only be predicted using existing layered reflection wave prediction techniques, inevitably affecting the accuracy of the detection. Summary of the Invention

[0006] In view of this, in order to at least partially solve the problem of the difficulty in detecting diverse oil and gas reservoirs in the existing technology, this invention proposes the following specific solutions for data processing and geological exploration of the reservoir space of diverse oil and gas reservoirs:

[0007] On the one hand, the present invention provides a method for detecting hidden geological anomalies based on wavefield decomposition, the method comprising:

[0008] S1. Perform random noise suppression preprocessing on the input dataset spin(⊿x,⊿y,⊿t) to be decomposed into wavefield, to obtain the original dataset sp(⊿x,⊿y,⊿t) with high signal-to-noise ratio;

[0009] S2. Spatial domain thinning is performed on the original dataset sp(⊿x,⊿y,⊿t) to obtain n common offset data subsets spn(n⊿x,n⊿y,⊿t), where n is a positive integer greater than 1;

[0010] S3. Perform wave field decomposition on each of the common offset data subsets spn(n⊿x,n⊿y,⊿t) to obtain n common offset data subsets spHn(n⊿x,n⊿y,⊿t) of normal field (strate reflection wave) with offset (n⊿x,n⊿y);

[0011] S4. Merge the n common offset data subsets spHn(n⊿x,n⊿y,⊿t) of normal field (strate reflection wave) into the original offset normal field (strate reflection wave) dataset spH(⊿x,⊿y,⊿t);

[0012] S5. Subtract the original offset normal field dataset spH(⊿x,⊿y,⊿t) from the original dataset sp(⊿x,⊿y,⊿t) to obtain the original offset anomalous field (diffraction wave, scattering wave) dataset spR(⊿x,⊿y,⊿t);

[0013] S6. Output the original dataset sp(⊿x,⊿y,⊿t), the original offset normal field dataset spH(⊿x,⊿y,⊿t), and the original offset anomaly field dataset spR(⊿x,⊿y,⊿t) to complete the wavefield decomposition process.

[0014] Preferably, in step S2, during the spatial domain thinning process, the trace spacing of the common offset data subset spn(n⊿x,n⊿y,⊿t) is (n⊿x,n⊿y), and n⊿x or n⊿y is not less than one-quarter of the wavelength of seismic waves propagating in the surrounding rock of the geological anomaly.

[0015] Preferably, in the common offset data subset spn(n⊿x,n⊿y,⊿t), within the diffraction wave distribution range with a limited lateral scale, only 1 / n channels of data are retained.

[0016] Preferably, in S3, one of the methods of wave field decomposition is S31 sparse-constrained Radon transform:

[0017] Using machine learning and applying the sparse-constrained Radon transform, a layered wavefield (normal field) data subset spHn(n⊿x,n⊿y,⊿t) is separated from the original wavefield data subset spn(n⊿x,n⊿y,⊿t); the objective function of machine learning is:

[0018] J = ||spn-LspHn|| 2 +α|spHn| p (0 < p ≤ 1)

[0019] In the formula, J is the objective function, L is the Radon inverse transform operator, spn is the high signal-to-noise ratio original data subset, spHn is the layered wave field data subset of the Radon domain prediction results, α is the sparsity constraint factor, and p is the Radon domain energy cluster focusing factor.

[0020] Preferably, for the objective function of machine learning, after minimization, the common offset data subset spHn(n⊿x,n⊿y,⊿t) of the normal field of the layered reflected wave can be solved.

[0021] Preferably, in step S3, the second method of wave field decomposition is S32 frequency spatial domain prediction and inversion:

[0022] Taking advantage of the predictability of regular wave fields in the frequency space domain, a predictive inversion technique is used to obtain a subset of common offset data of the normal field of layered reflected waves, spHn(n⊿x,n⊿y,⊿t).

[0023] The specific method for prediction and inversion in the frequency spatial domain is as follows:

[0024] S32-1. Set the objective function for prediction and inversion;

[0025] S32-2. Minimize the prediction inversion objective function to obtain the inversion equation;

[0026] S32-3. Perform prediction inversion in the frequency space domain, solve the inversion equation, and obtain the common offset data subset spHn(n⊿x,n⊿y,⊿t) of the normal field of the layered reflected wave.

[0027] Preferably, the inversion prediction objective function is:

[0028]

[0029] Where P is the prediction operator; spn is the input subset of the original wavefield data, which is a frequency slice; spHn is the subset of the regular wavefield data to be estimated; and λ is the control parameter.

[0030] Preferably, in step S32-2, the prediction inversion objective function is minimized to obtain the prediction inversion equation as follows:

[0031] [P H P+(λ+μ)I]spHn=λspn

[0032] In the formula, P is the prediction operator, P H is the conjugate of the prediction operator; spn(n⊿x,n⊿y,⊿t) is a subset of the input data, which is a frequency slice; spHn(n⊿x,n⊿y,⊿t) is the regular wave field signal (normal field) to be estimated; I represents the identity matrix; λ represents the control parameter.

[0033] Preferably, the prediction operator P is:

[0034]

[0035] Wherein, the operator length L is the number of sampling points in the seismic gather data.

[0036] Secondly, the present invention also provides a system for detecting hidden geological anomalies based on wavefield decomposition, the system comprising:

[0037] The preprocessing module performs random noise suppression preprocessing on the input dataset spin(⊿x,⊿y,⊿t) to be decomposed into wavefield, resulting in a high signal-to-noise ratio original dataset sp(⊿x,⊿y,⊿t).

[0038] The thinning processing module performs spatial domain thinning processing on the original dataset sp(⊿x,⊿y,⊿t) to obtain n common offset data subsets spn(n⊿x,n⊿y,⊿t), where n is a positive integer greater than 1;

[0039] The wave field decomposition module performs wave field decomposition on each of the co-offset data subsets spn(n⊿x,n⊿y,⊿t) to obtain n normal field (strate reflection wave) co-offset data subsets spHn(n⊿x,n⊿y,⊿t) with offsets of (n⊿x,n⊿y);

[0040] The merging module merges n common offset data subsets spHn(n⊿x,n⊿y,⊿t) of normal field (strate reflection wave) into the original offset normal field (strate reflection wave) dataset spH(⊿x,⊿y,⊿t);

[0041] The pruning module subtracts the original offset normal field dataset spH(⊿x,⊿y,⊿t) from the original dataset sp(⊿x,⊿y,⊿t) to obtain the original offset anomalous field (diffraction wave, scattered wave) dataset spR(⊿x,⊿y,⊿t);

[0042] The output module is used to output the original dataset sp(⊿x,⊿y,⊿t), the original offset normal field dataset spH(⊿x,⊿y,⊿t), and the original offset anomaly field dataset spR(⊿x,⊿y,⊿t) to complete the wavefield decomposition process.

[0043] Preferably, in the spatial domain thinning processing module, the trace spacing of the common offset data subset spn(n⊿x,n⊿y,⊿t) is (n⊿x,n⊿y), and n⊿x or n⊿y is not less than one-quarter of the wavelength of seismic waves propagating in the surrounding rock of the geological anomaly.

[0044] Preferably, in the common offset data subset spn(n⊿x,n⊿y,⊿t), within the diffraction wave distribution range with a limited lateral scale, only 1 / n channels of data are retained.

[0045] Thirdly, the present invention also provides a computer-readable medium storing instructions that can be read and executed by a computer, wherein the instructions, when executed, implement the method for detecting hidden geological anomalies based on wave field decomposition as described above.

[0046] Compared with the prior art, the technical solution of the present invention has the following beneficial effects: First, the solution is based on the theory of macroscopic scattering wave field (generalized diffraction wave field), and takes the difference in morphological characteristics between reflected waves and diffraction (scattering) wave fields as the theoretical basis for wave field decomposition; Second, the diffraction (scattering) wave field generated by irregular geological bodies with limited lateral scale in the comprehensive reflection wave field is randomized, and the diffraction (scattering) wave field with limited lateral scale is converted into a random wave field, while the reflection wave field with infinite lateral scale maintains the characteristics of a regular wave field; Third, a technology for decomposing regular wave fields and random wave fields has been developed, including: (1) a separation technology of layered wave fields and random wave fields and noise based on graph-guided sparse constraint Radon transform; (2) a separation technology of layered wave fields and random wave fields based on frequency spatial domain (fxy) prediction inversion; Finally, the observed comprehensive wave field is decomposed into the reflection wave field of layered medium and the diffraction (scattering) wave field generated by irregular geological bodies with limited lateral scale.

[0047] This solution addresses the problem of detailed description and characterization of the wavefield of irregular geological bodies such as fractured bodies, fractured-dissolved bodies, and fractured-cavitary bodies. Furthermore, it enables direct detection of small-scale, irregular, and concealed oil and gas reservoirs hidden within layered wavefields, making it an effective solution to the challenges of direct detection of the aforementioned types of oil and gas reservoirs. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1a This is an irregular volume model with finite lateral dimensions, as described in this embodiment of the invention.

[0050] Figure 1b The diffracted wave formed by a geological body with a finite lateral scale in this embodiment of the invention and its offset-processed "short-axis reflected wave";

[0051] Figure 1c This is a grayscale view of a geological body offset profile with a limited lateral scale according to an embodiment of the present invention.

[0052] Figure 2 This invention illustrates the effect of lateral scale variation of geological bodies on the characteristics of short-axis reflected waves after offset.

[0053] Figure 3a This is a model of an irregular geological body beneath a non-smooth weathered surface, as described in an embodiment of the present invention.

[0054] Figure 3bThe diffracted wave and the offset short-axis reflection of the irregular geological body under the non-smooth weathered surface in this embodiment of the invention have a wavelet frequency of 30Hz.

[0055] Figure 3c This is a short-axis reflection of a diffracted wave from an irregular geological body with a width of 60m and a height of 0-40m beneath a rough weathered surface, according to an embodiment of the present invention.

[0056] Figure 3d This is a short-axis reflection of a diffracted wave from an irregular geological body with a height of 6m and a width of 10-300m beneath a rough weathered surface, according to an embodiment of the present invention.

[0057] Figure 4 The diffraction wave amplitude curves of an irregular body with a height of 15m and widths of 600, 300, 200, 100, 70, 50, 30, and 15m are shown in the embodiments of the present invention.

[0058] Figure 5a Examples of cave models filled with different materials according to embodiments of the present invention;

[0059] Figure 5b This is a property diagram of the cave filling material according to an embodiment of the present invention;

[0060] Figure 5c The difference in short-axis reflection of diffracted (scattered) waves after deflection in different filling cavities according to embodiments of the present invention;

[0061] Figure 6 This is an example of how graphic guided wave field decomposition is used to separate layered wave fields from random wave fields (noise) according to an embodiment of the present invention.

[0062] Figure 7 The following is a simulation diagram of the normal and anomalous fields obtained by predicting the deconvolution in the fxy domain according to an embodiment of the present invention; where a is the horizontally extended regular wave field, b is the local anomalous map, c is the combined wave field map of a and b, and d is the wave field map obtained by predicting the deconvolution in the fxy domain.

[0063] Figure 8 This is a flowchart of the wave field decomposition process according to an embodiment of the present invention;

[0064] Figure 9 The coherent property volume after wave field decomposition of T9b reflected wave from a weathered Ordovician surface in a basin according to an embodiment of the present invention demonstrates carbonate karst landforms.

[0065] Figure 10 This invention relates to sinkholes and their distribution that appeared after wave field decomposition of the Ordovician subsurface in a basin. Detailed Implementation

[0066] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be understood that the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0067] Those skilled in the art should understand that the following specific embodiments or implementation methods are a series of optimized configurations listed to further explain the specific content of the invention. These configurations can be combined or used in conjunction with each other, unless the invention explicitly states that some or a specific embodiment or implementation method cannot be associated with or used in conjunction with other embodiments or implementation methods. Furthermore, the following specific embodiments or implementation methods are merely optimized configurations and are not intended to limit the scope of protection of the invention.

[0068] The present invention will now be described in detail with reference to the accompanying drawings. To address the current technical problems, the technical solution of the present invention mainly includes the following three aspects:

[0069] The invention comprises wavefield decomposition based on the theory of macroscopic scattered wavefields (generalized diffracted wavefields); randomization of diffracted (scattered) wave morphology with selective randomization of wavefield data as the key; and wavefield decomposition centered on the difference between regular and random wavefield morphological characteristics. The following description details the solutions proposed in this invention.

[0070] I. Wavefield Decomposition Based on Macroscopic Scattered Wavefield (Generalized Diffraction Wavefield) Theory

[0071] Macroscopic scattered wave fields possess the following fundamental characteristics:

[0072] (1) The seismic wave field is the scattered wave field, which is the secondary wave generated when the seismic wave encounters an obstacle in the propagation path;

[0073] (2) The necessary and sufficient condition for the generation of scattered waves (i.e., secondary waves, the same below) is the discontinuity caused by an obstacle located on the propagation path of the incident wave. Physically, the surface of the obstacle is a discontinuity (discontinuity) of the medium parameters.

[0074] (3) Whether the scattered wave is generated and its distribution pattern are also related to the geometric shape and spatial scale (linearity) of the obstacle. According to the size of the geological body that forms the scattered wave field (especially the lateral scale), the generated waves are divided into reflected waves, diffracted waves and scattered waves.

[0075] (4) The amplitude of the scattered wave is related to the lateral scale of the geological body and also to the contrast (difference) of the medium parameters on both sides of the boundary: the larger the lateral scale, the greater the contrast, and the stronger the secondary wave.

[0076] In the above-mentioned item (3), the wave fields generated, depending on the size of the geological body forming the scattered wave field (especially the lateral scale), are reflected waves, diffracted waves, and scattered waves (scattered waves in the narrow sense, the same below). These three types of waves have different morphological characteristics, including:

[0077] ① A layered medium with a large lateral scale that extends infinitely produces reflected waves, which have a regular shape characteristic of extending infinitely laterally.

[0078] ② Irregular geological bodies with small lateral scales generate diffracted (scattered) waves, exhibiting regular morphological characteristics with limited lateral scales.

[0079] The morphological characteristics of the reflected waves, diffracted waves, and scattered waves will change before and after the data offset processing. These changes include:

[0080] ① Before and after data offset processing, the undulation structure of the reflected wave field changes to some extent, but still maintains the regular characteristic of infinite horizontal extension;

[0081] ② Before data offset processing, diffracted waves and scattered waves have regular morphological characteristics of quadratic curves (surfaces), but their lateral scale is limited; after data offset processing, the lateral scale of diffracted waves and scattered waves is further reduced, forming a beaded pattern of "short axis + tail", but still retaining regular morphological characteristics.

[0082] ③ Before and after data offset processing, reflected waves, diffracted waves and scattered waves all have regular morphological characteristics, but the lateral extension scale of reflected waves is "infinite", while the lateral extension scale of diffracted waves and scattered waves is "limited".

[0083] To investigate the influence of the scale (linear dimensions) and geometry of geological bodies on the variation of wavefield characteristics, we conducted a study using elastic wave forward modeling of geological bodies of arbitrary shapes with finite lateral scales. This yielded the variation characteristics of the diffracted (scattered) wavefield morphology in self-excited and self-recovered synthetic seismic records, including:

[0084] ① Wavefield morphology characteristics of irregular geological bodies

[0085] Figure 1a This is an example of an irregular geological body model with a limited lateral scale. The P-wave velocity of the surrounding rock of the irregular geological body is 5500 m / s, the dominant frequency of the reflected wavelet is 20 Hz, the wavelength is 275 m, the quarter wavelength in the surrounding rock medium is 68.75 m, the CDP point spacing in the profile is 5 m, and 65 m in the profile is approximately equivalent to a quarter wavelength in the surrounding rock medium. Figure 1b For seismic wave diffractions (scattering) from irregular geological bodies with finite lateral scale and arbitrary shapes, the morphological characteristics before migration are those of a quadratic curve (surface) with finite lateral scale. After migration, the characteristics of short reflections with a "short axis + tail" are observed (e.g., Figure 1bAs shown in the diagram above, the "short axis" is the main body of the diffracted (scattered) wave energy. The tail is formed by the offset algorithm, and its amplitude is weak, gradually weakening and disappearing towards both sides. The main body of the "short axis" forms a beaded shape on the gray cross-section, such as... Figure 1c As shown, this is a quasi-regular form with a finite scale.

[0086] Both un-misaligned and misaligned waves have regular morphological characteristics with limited lateral scale.

[0087] ② Relationship between the lateral scale of irregular geological bodies and the lateral extension scale of the wavefield

[0088] Figure 2 These are the short-axis "reflected waves" ("short axis + tail") of diffracted waves from self-excited and self-receiving profiles of geological bodies with transverse scales of 10, 30, and 100 m, after being processed by migration. If we take the point where the slope of the amplitude change between the short axis and the tail on the self-excited and self-receiving profile is the maximum as the length of the short axis body, we can see that the length of the short axis body and the width of the irregular body have the following relationship:

[0089] When the lateral scale of a geological body is larger than a quarter wavelength in the surrounding rock, the length of the short axis is proportional to the lateral scale of the geological body. When the lateral scale of a geological body is smaller than a quarter wavelength in the surrounding rock, the length of the short axis remains at 65m, meaning that the quarter wavelength in the surrounding rock remains basically unchanged, and only the amplitude of the short axis decreases as the width decreases.

[0090] In the model, the P-wave velocity of the surrounding rock of the geological body is 5500 m / s, the dominant frequency of the reflected wavelet is 20 Hz, the wavelength is 275 m, the quarter wavelength is 68.75 m, and the CDP point spacing is 5 m. Therefore, a length of 65 m is approximately equivalent to one-quarter of the wavelength in the surrounding rock. In other words, as long as the lateral scale of the geological body is not greater than one-quarter of the wavelength of the surrounding rock, the length of the minor axis of its diffracted (scattered) waves after migration processing will roughly maintain one-quarter of the wavelength of the surrounding rock. This is the limit of the lateral resolution of the seismic wavefield after migration processing. Conversely, only when the lateral scale of the geological body is not less than one-quarter of the wavelength of the surrounding rock will the length of the minor axis after migration processing be proportional to the lateral scale of the geological body, and can be used to determine the width of the geological body.

[0091] It is evident that the diffracted (scattered) wave field of a geological body with a limited lateral scale exhibits regular morphological characteristics with a limited lateral scale.

[0092] ③ Morphological characteristics of wave fields of geological bodies at different longitudinal and transverse scales

[0093] Figure 3a This is a model of an irregular geological body beneath a rough weathered surface. The incident wavelet frequency is 30Hz, the longitudinal wave wavelength within the geological body is approximately 47m, and the longitudinal wave wavelength in the surrounding rock of the geological body is 200m.

[0094] Figure 3b The self-excited and self-absorbed synthetic seismic profiles of single geological body models with the same height (6m) but different widths (60m and 20m), as well as diffracted waves and short-axis reflected waves after offset on the offset profiles.

[0095] Figure 3c Forward migration profiles of geological body models with an average width of 60m and average heights of 2, 4, 6, 8, 10, 12, 14, 16, 20, 25, 30, 35, and 40m.

[0096] Figure 3d Forward migration profiles of geological body models with an average height of 6m and average widths of 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 120, 180, 240, and 300m respectively.

[0097] The forward modeling results in the above example show that the wavefield of irregular geological bodies exhibits beaded characteristics in the short-axis reflections after migration.

[0098] For geological bodies with the same width but different heights, when the height is less than 25m (equivalent to half the wavelength within the geological body), the shape of the beads does not change much. When the height of the geological body is greater than 25m, the top and bottom surfaces reflect and separate. Due to the transmission loss at the interface and the influence of multiple reflections from the top and bottom surfaces, the lower half of the beads changes, but still appears as beads.

[0099] For geological bodies with the same height but different widths, the amplitude of the beads varies with the width; when the width of the geological body is no more than 100m, the width of the beads remains basically unchanged; when the width of the geological body is greater than 100m, the width of the beads increases with the increase of the width.

[0100] ④ Characteristics of amplitude variation of reflected waves, diffracted waves, and scattered waves of geological bodies at different lateral scales

[0101] Geological bodies of the same height (e.g., 15m) but different widths will produce reflected waves, diffracted waves, and scattered waves, such as... Figure 4 As shown. When the width is very small, such as 15m, the amplitude of the secondary wave it produces is weak, but the extreme value range is wide, exhibiting scattering characteristics; when the width is medium, such as 50-100m, the amplitude of the main extreme value increases significantly with the increase of the width, exhibiting diffraction wave characteristics; when the width is 200-600m, the extreme value region of the amplitude curve presents a fluctuating plateau, exhibiting reflected wave + edge diffraction wave characteristics.

[0102] Figure 4The results show that the wave field of different geological bodies varies with the lateral scale; as the lateral scale increases, the wave gradually changes from scattered waves to diffracted waves and reflected waves; as the lateral scale increases, the wave amplitude gradually increases. This characteristic indicates that, under the condition that the medium properties remain unchanged, the width and amplitude decrease, which is an important reason why diffracted and scattered waves are often submerged in layered reflections.

[0103] ⑤ The influence of differences in the physical properties of irregular geological bodies on the characteristics of diffracted (or scattered) wave fields.

[0104] Figure 5a For cave model, Figure 5b This is a list of physical properties of different filling materials in a cave. The interface reflection coefficients of the cave for different filling materials range from 0.07 to 0.89. The short-axis reflections of the diffracted (scattered) waves after deflection are shown for caves with different interface reflection coefficients. Figure 5c As can be seen, the resulting differences are: a large interface reflection coefficient results in strong short-axis reflection, and vice versa.

[0105] II. Wavefield Decomposition Based on the Differences in Lateral Extension Morphology between Layered and Random Wavefields

[0106] The above analysis shows that the layered reflection wave field has a morphological feature of "infinite" lateral extension, while diffracted waves and scattered waves have a morphological feature of "limited" lateral extension. This difference in the morphological features of the two types of wave fields lays the foundation for the selective randomization of wave field data.

[0107] The selective randomization of the wavefield data is based on the differences in morphological characteristics between reflected waves and diffracted (scattered) waves: regardless of whether the reflected waves have undergone migration processing, they are regular wavefields (normal fields) with "infinite" lateral extension; diffracted and scattered waves without migration processing are regular wavefields with quadratic curve (surface) morphological characteristics, with a width greater than the geological body but a limited lateral scale; diffracted and scattered waves after migration processing are regular wavefields with a width at least one-quarter of the seismic wavelength in the surrounding rock, characterized by a beaded pattern of "short axis + tail".

[0108] For example, in the specific implementation of selective randomization processing of wavefield data, the present invention uses a spatial domain thinning sampling interval method for the comprehensive wavefield composed of reflected waves and diffracted (scattered) waves. When the selected thinning sampling interval is appropriate, the horizontally extended "infinite" reflected wavefield can maintain the regular shape characteristics of "infinite" extension, while the horizontally finite diffracted (scattered) wavefield can be transformed into random shape characteristics.

[0109] For example, the selective randomization processing of the wavefield data can be implemented by methods that include:

[0110] ① Set the data to be decomposed into a wavefield as the original dataset sp(⊿x,⊿y,⊿t) with high signal-to-noise ratio and inter-channel offset of (⊿x,⊿y), where ⊿t is the time sampling interval;

[0111] ② The dataset sp(⊿x,⊿y,⊿t) is thinned in the spatial domain to obtain n common offset data subsets spn(n⊿x,n⊿y,⊿t) with offsets of (n⊿x,n⊿y). At this time, the trace spacing of the data subset becomes (n⊿x,n⊿y), where n is an integer value greater than 1, and n⊿x or n⊿y is not less than one-quarter of the wavelength of seismic waves propagating in the surrounding rock of the geological anomaly.

[0112] ③ In the common offset data subset spn(n⊿x,n⊿y,⊿t), within the range of diffracted (scattered) wave distribution with limited lateral scale, only 1 / n channels of data are retained.

[0113] ④ The choice of the value n can be determined by comparing multiple parameters and based on the randomization effect of the diffracted (scattered) wave.

[0114] III. Two wavefield decomposition methods based on morphological differences

[0115] Wave field decomposition can be achieved using two methods, including:

[0116] Method ①: A method for separating layered wavefields from random wavefields and noise based on machine learning sparse-constrained Radon transform.

[0117] We can apply the idea of ​​graph-guided transformation to the frequency domain parabolic Radon transform based on machine learning. Under sparse constraints, the energy clusters in the slow domain are more convergent and focused, with high resolution, making it easier to distinguish between horizontally "infinitely" extending wave fields and randomly distributed wave fields. This allows us to separate regularly shaped layered wave fields from randomly distributed noise.

[0118] Preferably, the objective function in machine learning is:

[0119] J = ||spn-LspHn|| 2 +α|spHn| p (0 < p ≤ 1)

[0120] In the formula, J is the objective function, L is the Radon inverse transform operator, spn is the original trace data subset, spHn is the layered wave field data subset of the Radon domain prediction result, α is the sparsity constraint factor, the larger the factor, the sparser the result, and p is the Radon domain energy cluster focusing factor, the smaller p is, the more focused.

[0121] By using machine learning, we can obtain a reasonable subset of spHn data from historical data. The separation effect of this method is as follows: Figure 6As shown, the left image is the composite wave field before separation, the middle image is the regular wave field after separation, and the right image is the random noise wave field after separation.

[0122] Method ②: A method for separating layered wavefields and random wavefields based on frequency spatial domain (fxy domain) prediction and inversion.

[0123] A regular wave field in the spatiotemporal domain can be approximated by superimposing a finite number of plane wave signals in that domain. The superposition result of a finite number of plane waves has a linear predictive relationship in the frequency space domain (i.e., the fxy domain). The prediction filtering error reflects that there are also irregular plane waves (i.e., irregular random wave fields) in the space of the signal.

[0124] The laterally extended regular wave field is extracted by prediction and inversion, and the difference between it and the original wave field is obtained, which is the randomly distributed irregular wave field.

[0125] We set the objective function as follows:

[0126]

[0127] In the formula, P is the prediction operator; spn is the input subset of the original wavefield data, which is a frequency slice; spHn is the subset of the regular wavefield data to be estimated; and λ is the control parameter.

[0128] The inversion equation is

[0129] [P * P+(λ+μ)I]spHn=λspn

[0130] Where P is the prediction operator, P * For the conjugate of the prediction operator, the operator length L is the number of sampling points in the seismic gather data; spn is the subset of input data, which is a frequency slice; spHn is the regular wavefield signal (normal field) to be estimated; I represents the identity matrix; λ represents the control parameter; μ is a parameter that determines the degree of energy retention of the regular wavefield. The larger μ is, the more energy of the regular wavefield is retained; conversely, the smaller μ is, the less energy of the regular wavefield is retained.

[0131] In a more specific implementation, the prediction operator P can take the following form:

[0132]

[0133] Wherein, the operator length L is the number of sampling points in the seismic gather data.

[0134] Figure 7 This example demonstrates a simulation study for predicting and resolving normal and anomalous fields in the frequency spatial domain (fxy domain). The normal field is the reflected wave from the layered medium. Figure 7a) The anomalous field is designed so that the point-like diffracted (scattered) wave migration under extreme conditions converges into a point-like anomalous wave. Figure 7 b), the two superimposed form a composite wave field ( Figure 7 c), after fxy domain prediction and deconvolution processing, point-like anomalous waves are obtained. Figure 7 d). By Figure 7 It can be seen that the method for separating the layered wave field and the random wave field by using the fxy domain prediction and inversion proposed in this embodiment can effectively separate the anomalous waves.

[0135] Furthermore, combining Figure 8 The diagram shown is a detailed flowchart of the method of the present invention. The wavefield decomposition method based on morphological feature differences described in the present invention preferably includes the following steps:

[0136] Step 1: Data preparation: The original dataset spin(⊿x,⊿y,⊿t) to be decomposed into wavefield is preprocessed to suppress random noise, resulting in a high signal-to-noise ratio original dataset sp(⊿x,⊿y,⊿t).

[0137] The second step is to randomize (i.e., thin out) the original high signal-to-noise ratio dataset sp(⊿x,⊿y,⊿t) to obtain n common offset data subsets spn(n⊿x,n⊿y,⊿t) with offsets of (n⊿x,n⊿y).

[0138] The third step is to process each thinned common offset data subset spn(n⊿x,n⊿y,⊿t) using one of the two wave field decomposition techniques described above in this embodiment, namely, wave field decomposition by combining sparse constraint Radon transform with machine learning, or wave field decomposition by using frequency spatial domain inversion.

[0139] Fourth step: After processing with wave field decomposition technology, n subsets of normal field co-offset data of layered reflected waves with offset distances of (n⊿x, n⊿y) are obtained;

[0140] Step 5: Merge the n offset normal field co-offset data subsets spHn((n⊿x,n⊿y,⊿t) of the layered reflected wave normal field with offset distances of (n⊿x,n⊿y) into the original offset normal field dataset spH(⊿x,⊿y,⊿t);

[0141] Step 6: Subtract the original offset normal field dataset spH(⊿x,⊿y,⊿t) from the original high signal-to-noise ratio reflected wave dataset sp(⊿x,⊿y,⊿t) to obtain the original offset anomaly field dataset spR(⊿x,⊿y,⊿t) which only has random distribution characteristics;

[0142] Step 7: Output the original reflected wave dataset sp(⊿x,⊿y,⊿t), the original offset normal field dataset spH(⊿x,⊿y,⊿t), and the original offset anomaly field dataset spR(⊿x,⊿y,⊿t) to complete the wavefield decomposition process. Among them, spH(⊿x,⊿y,⊿t) is the reflected wave dataset of the layered medium, and spR(⊿x,⊿y,⊿t) is the diffracted (scattered) wave dataset of irregular geological bodies with limited lateral scale.

[0143] Below, we verify the actual effectiveness of this scheme based on wavefield data from actual geological exploration data.

[0144] Figure 9 This is an example of wave field decomposition of a Paleozoic Ordovician carbonate reservoir (body) in a basin. The weathered surface and the multiple coal seams of varying thicknesses above it create strong reflections (such as...). Figure 9 As shown above, strong reflections, besides minor fluctuations, show few other characteristics; after wave field decomposition, strong reflections are suppressed into weak reflections, resulting in multiple "beads" formed by diffracted (scattered) waves (such as...). Figure 9 (As shown in the middle figure); coherence properties were extracted from the three-dimensional diffracted (scattered) wave field data to obtain T9b slices reflecting the weathering surface morphology of Lower Ordovician carbonate rocks (as shown in the middle figure); Figure 9 As shown below, the coherence properties demonstrate the changes in the structure of the weathered surface, which includes karst peak forests, peak forest valleys, karst depressions and plains, as well as river channels.

[0145] Figure 10 This is another geological phenomenon on the Ordovician weathering surface of a certain basin: a karst sinkhole. In the anomalous field after wave field decomposition, a string of beads, larger at the top and smaller at the bottom, appears on the cross-section; a strong-amplitude underground karst sinkhole appears on the plane, with a mouth width of 800-950 meters and a karst depth of 420 meters. On a larger scale, the spatial distribution characteristics of the karst sinkhole can also be seen, extending longitudinally from the Ordovician weathering surface to the underlying Ma Wudi. Figure 10 In the image, the left image is a seismic profile before wavefield decomposition, the middle image is a seismic profile after wavefield decomposition, and the right image is a map showing the distribution of sinkholes on the slice after wavefield decomposition.

[0146] In yet another embodiment, the solution of the present invention can also be implemented by a concealed geological anomaly detection system based on wavefield decomposition, the system comprising:

[0147] The preprocessing module performs random noise suppression preprocessing on the input dataset spin(⊿x,⊿y,⊿t) to be decomposed into wavefield, resulting in a high signal-to-noise ratio original dataset sp(⊿x,⊿y,⊿t).

[0148] The thinning processing module performs spatial domain thinning processing on the original dataset sp(⊿x,⊿y,⊿t) to obtain n common offset data subsets spn(n⊿x,n⊿y,⊿t), where n is a positive integer greater than 1;

[0149] The wave field decomposition module performs wave field decomposition on each of the co-offset data subsets spn(n⊿x,n⊿y,⊿t) to obtain n normal field (strate reflection wave) co-offset data subsets spHn(n⊿x,n⊿y,⊿t) with offsets of (n⊿x,n⊿y);

[0150] The merging module merges n common offset data subsets spHn(n⊿x,n⊿y,⊿t) of normal field (strate reflection wave) into the original offset normal field (strate reflection wave) dataset spH(⊿x,⊿y,⊿t);

[0151] The pruning module subtracts the original offset normal field dataset spH(⊿x,⊿y,⊿t) from the original dataset sp(⊿x,⊿y,⊿t) to obtain the original offset anomalous field (diffraction wave, scattered wave) dataset spR(⊿x,⊿y,⊿t);

[0152] The output module is used to output the original dataset sp(⊿x,⊿y,⊿t), the original offset normal field dataset spH(⊿x,⊿y,⊿t), and the original offset anomaly field dataset spR(⊿x,⊿y,⊿t) to complete the wavefield decomposition process.

[0153] Furthermore, in the thinning processing module, the trace spacing of the common offset data subset spn(n⊿x,n⊿y,⊿t) is (n⊿x,n⊿y), and n⊿x or n⊿y is not less than one-quarter of the wavelength of seismic waves propagating in the surrounding rock of the geological anomaly.

[0154] Furthermore, within the co-offset data subset spn(n⊿x,n⊿y,⊿t), only 1 / n channels of data are retained within the diffraction wave distribution range with a limited lateral scale.

[0155] In another embodiment, this solution can be implemented using a device, which may include corresponding modules that perform one or more steps in the various embodiments described above. Therefore, each or more steps in the various embodiments can be performed by a corresponding module, and the electronic device may include one or more of these modules. A module may be one or more hardware modules specifically configured to perform a corresponding step, or implemented by a processor configured to perform a corresponding step, or stored in a computer-readable medium for implementation by a processor, or implemented through some combination thereof.

[0156] This device can be implemented using a bus architecture. A bus architecture can include any number of interconnect buses and bridges, depending on the specific application of the hardware and overall design constraints. The bus connects various circuits, including one or more processors, memory, and / or hardware modules. The bus can also connect various other circuits such as peripherals, voltage regulators, power management circuitry, external antennas, etc.

[0157] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of this solution includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which the embodiments of this solution pertain. The processor performs the various methods and processes described above. For example, the method embodiments of this solution can be implemented as software programs tangibly contained in a machine-readable medium, such as memory. In some embodiments, part or all of the software program can be loaded and / or installed via memory and / or a communication interface. When the software program is loaded into memory and executed by the processor, one or more steps of the methods described above can be performed. Alternatively, in other embodiments, the processor can be configured to perform one of the methods described above by any other suitable means (e.g., by means of firmware).

[0158] The logic and / or steps represented in the flowchart or otherwise described herein may be specifically implemented in any readable storage medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0159] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for detecting hidden geological anomalies based on wavefield decomposition, characterized in that, The method includes: S1. Preprocess the input dataset spin(⊿x,⊿y,⊿t) to be decomposed into wavefield by suppressing random noise to obtain the original dataset sp(⊿x,⊿y,⊿t) with high signal-to-noise ratio; where (⊿x,⊿y) is the inter-channel offset and ⊿t is the time sampling interval. S2. Spatial domain thinning is performed on the original dataset sp(⊿x,⊿y,⊿t) to obtain n common offset data subsets spn(n⊿x, n⊿y,⊿t), where n is a positive integer greater than 1; S3. Perform wave field decomposition on each of the co-offset data subsets spn(n⊿x, n⊿y, ⊿t) to obtain n normal field co-offset data subsets spHn(n⊿x, n⊿y, ⊿t) with offsets of (n⊿x, n⊿y); S4. Merge the n normal field co-offset data subsets spHn(n⊿x, n⊿y, ⊿t) into the original offset normal field dataset spH(⊿x, ⊿y, ⊿t); S5. Subtract the original offset normal field dataset spH(⊿x,⊿y,⊿t) from the original dataset sp(⊿x,⊿y,⊿t) to obtain the original offset anomaly field dataset spR(⊿x,⊿y,⊿t); S6. Output the original dataset sp(⊿x,⊿y,⊿t), the original offset normal field dataset spH(⊿x,⊿y,⊿t), and the original offset anomaly field dataset spR(⊿x,⊿y,⊿t) to complete the wavefield decomposition process.

2. The method according to claim 1, characterized in that, In S2, during the spatial domain thinning process, the trace spacing of the common offset data subset spn(n⊿x, n⊿y, ⊿t) is (n⊿x, n⊿y), and n⊿x or n⊿y is not less than one-quarter of the wavelength of seismic waves propagating in the surrounding rock of the geological anomaly.

3. The method according to claim 2, characterized in that, In the common offset data subset spn(n⊿x , n⊿y,⊿t), within the diffraction wave distribution range with limited lateral scale, only 1 / n channels of data are retained.

4. The method according to claim 1, characterized in that, In S3, the wave field decomposition method is the sparse-constrained Radon transform method: By using sparse-constrained Radon transform combined with machine learning, a layered wavefield data subset spHn(n⊿x, n⊿y, ⊿t) is separated from the original wavefield data subset spn(n⊿x, n⊿y, ⊿t).

5. The method according to claim 4, characterized in that, The objective function of the sparse-constrained Radon transform is: ， Where J is the objective function, L is the Radon inverse transform operator, spn is the original data subset, spHn is the layered wave field data subset of the Radon domain prediction results, α is the sparsity constraint factor, and p is the Radon domain energy cluster focusing factor.

6. The method according to claim 1, characterized in that, In S3, the wave field decomposition method is a frequency spatial domain prediction-inversion wave field decomposition method: leveraging the predictability of regular wave fields in the frequency spatial domain, prediction-inversion techniques are used to separate the normal field of layered reflected waves. S32-1. Set the objective function for prediction and inversion; S32-2. Minimize the prediction inversion objective function; S32-3. After minimization, the frequency spatial domain prediction inversion equation is obtained. Solve the prediction inversion equation to obtain the common offset data subset spHn(n⊿x,n⊿y,⊿t) of the normal field of the layered reflected wave.

7. The method according to claim 6, characterized in that, The objective function for prediction and inversion is: Where P is the prediction operator; spn is the high signal-to-noise ratio original wavefield data subset, which is a frequency slice; spHn is the regular wavefield data subset to be estimated; and λ is the control parameter.

8. The method according to claim 6, characterized in that, In S32-3, the prediction inversion equation is: Where P is the prediction operator, P H λ represents the conjugate of the prediction operator; spn represents the subset of input data, which is a frequency slice; spHn represents the regular wave field signal to be estimated; I represents the identity matrix; λ represents the control parameter.

9. A system for detecting hidden geological anomalies based on wavefield decomposition, characterized in that, The system includes: The preprocessing module is used to preprocess the input dataset spin(⊿x,⊿y,⊿t) to be decomposed into wavefield by random noise, so as to obtain the original dataset sp(⊿x,⊿y,⊿t) with high signal-to-noise ratio; where (⊿x,⊿y) is the inter-channel offset and ⊿t is the time sampling interval. The thinning processing module is used to perform spatial domain thinning processing on the original dataset sp(⊿x,⊿y,⊿t) to obtain n common offset data subsets spn(n⊿x,n⊿y,⊿t), where n is a positive integer greater than 1; The wave field decomposition module is used to perform wave field decomposition on each of the co-offset data subsets spn(n⊿x,n⊿y,⊿t) to obtain n normal field co-offset data subsets spHn(n⊿x,n⊿y,⊿t) with offsets of (n⊿x,n⊿y); The merging module is used to merge n normal field co-offset data subsets spHn(n⊿x,n⊿y,⊿t) into the original offset normal field dataset spH(⊿x,⊿y,⊿t); The pruning module is used to subtract the original offset normal field dataset spH(⊿x,⊿y,⊿t) from the original dataset sp(⊿x,⊿y,⊿t) to obtain the original offset anomaly field dataset spR(⊿x,⊿y,⊿t); The output module is used to output the original dataset sp(⊿x,⊿y,⊿t), the original offset normal field dataset spH(⊿x,⊿y,⊿t), and the original offset anomaly field dataset spR(⊿x,⊿y,⊿t) to complete the wavefield decomposition process.

10. A computer-readable medium, characterized in that, The medium stores instructions that can be read and executed by a computer, and when the instructions are executed, they implement the method for detecting hidden geological anomalies based on wave field decomposition as described in any one of claims 1-8.

Images