Carbonate karst cave artificial intelligence identification method and related device

Abstract

The carbonate karst cave artificial intelligence identification method comprises the steps that post-stack seismic inversion and instantaneous amplitude are subjected to same-level stacking to obtain karst cave distribution characteristics, and the size and the amplitude of the karst cave distribution characteristics are counted; obtaining an initial stratified model by utilizing seismic horizon and post-stack seismic inversion, and superposing the initial stratified model and the forward karst cave model to obtain a forward karst cave model based on a work area; performing convolution on the forward modeling karst cave model based on the work area and the seismic wavelets of the work area to synthesize a seismic record, taking the forward modeling karst cave model as a label, and training the seismic record by utilizing a CNN (Convolutional Neural Network); and finally, inputting actual seismic data into the trained neural network to obtain a finely depicted carbonate karst cave data body. The forward modeling model is used as a training label and participates in karst cave prediction under the CNN neural network method, so that the accuracy and credibility of karst cave identification are enhanced.

Description

Technical Field

[0001] The present invention belongs to the field of seismic reservoir prediction, and in particular relates to an artificial intelligence identification method for carbonate caves and related devices. Background Art

[0002] With the continuous deepening of oil and gas exploration and development, the detailed description of deep carbonate reservoirs is becoming more and more difficult. Cave carbonate reservoirs are a very important type of oil and gas reservoirs. More than half of the global oil and natural gas production is distributed in cave oil and gas reservoirs. Under such circumstances, it is very important to carry out research on caves and cave oil and gas reservoirs. If the spatial distribution of caves can be accurately predicted, it will play an important role in studying cave oil and gas reservoirs, improving the success rate of exploration, and improving development efficiency.

[0003] There are many commonly used cave prediction methods in the industry, including cumulative energy difference method, coherent attribute analysis, curvature attribute analysis method, seismic prestack data analysis technology, etc. However, these methods are all based on the "lateral discontinuity" of the characteristic amplitude of cave seismic response for detection. In practical applications, these methods are affected by other geological factors (such as faults, lithology and other discontinuous structures) and non-geological factors (such as noise, etc.), and cannot accurately reflect the location and geometric morphology of cave development. In recent years, machine learning, especially deep learning, has developed rapidly, making it possible to process big data in a structured manner and greatly improving the efficiency of data processing. At the same time, the excellent performance of convolutional neural networks in image classification and recognition has made them widely used in the field of seismic interpretation. Therefore, some scholars have used CNN neural network methods to identify and characterize heterogeneous carbonate cave reservoirs. However, the key problem of CNN neural networks is that they require massive seismic data with cave characteristics for training. In the actual exploration process, the accurately interpreted carbonate cave seismic data are very scarce, resulting in low accuracy and reliability of cave identification. Summary of the invention

[0004] The purpose of the present invention is to provide an artificial intelligence identification method and related devices for carbonate caves, so as to solve the problem that the accurately interpreted seismic data of carbonate caves are very scarce in the actual exploration process, resulting in low accuracy and credibility of cave identification.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, the present invention provides an artificial intelligence identification method for carbonate caves, comprising:

[0007] The distribution characteristics of the caves are obtained by superimposing the post-stack seismic inversion with the instantaneous amplitude at the same level, and their sizes and amplitudes are counted;

[0008] The initial layered model is obtained by using seismic horizons and post-stack seismic inversion, and the initial layered model is superimposed on the forward cave model to obtain a forward cave model based on the work area;

[0009] The forward-modeled karst cave model based on the work area is convolved with the seismic wavelet of the work area to synthesize the seismic records. The forward-modeled karst cave model is used as a label to train the seismic records using the CNN neural network.

[0010] Finally, the actual seismic data is input into the trained neural network to obtain a finely characterized carbonate cave data volume.

[0011] Furthermore, the post-stack seismic inversion and instantaneous amplitude are superimposed at the same level to obtain the distribution characteristics of the caves, and their size and amplitude are counted:

[0012] Based on the post-stack seismic inversion, the instantaneous amplitude of the seismic records is stacked at the same level to obtain the distribution of carbonate caves;

[0013] Statistical size of carbonate caves (r x ,r y ,r z ) and the amplitude anomaly size A, and the average value of the cave size and the earthquake amplitude anomaly (μ x ,μ y ,μ z ,μ A ), and standard deviation (σ x ,σ y ,σ z ,σ A );

[0014]

[0015]

[0016]

[0017]

[0018] Where μ x ,μ y ,μ z ,μ A are the average values ​​of cave size and earthquake amplitude, σ x , σ y , σ z , σ A are the standard deviations of cave size and earthquake amplitude, A i are the size of the ith cave and the earthquake amplitude, N is the number of caves, and x, y, and z are the east-west, north-south, and vertical directions, respectively.

[0019] Furthermore, the initial layered model is obtained by using seismic horizons and post-stack seismic inversion:

[0020] Under the constraint of seismic horizon, the average amplitude of the interlayer inversion results is calculated and set as the amplitude value of the layer segment to obtain the initial layered seismic model of the layered medium.

[0021] Furthermore, the initial layered model is superimposed on the forward cave model to obtain a forward cave model based on the work area, and a seismic model containing multiple cave sizes and amplitude characteristics is superimposed.

[0022] Furthermore, the main feature of caves in seismic data is the "beaded" structure, that is, the high amplitude value that is concave downward; in the seismic profile, each cave is characterized by a vertically elongated ellipsoid; inside the ellipsoid, the seismic waveform extends downward and shows a high amplitude value; the carbonate cave is forward modeled using the ellipsoid calculation formula:

[0023]

[0024] The spherical coordinates of the ellipsoid are (x, y, z), and the radii of the ellipsoid in three directions satisfy r x ∈[μ x -σ x ,μ x +σ x ],r y ∈[μ y -σ y ,μ y +σ y ],r z ∈[μ z -σ z ,μ z +σ z ], and the amplitude inside the ellipsoid should satisfy A∈[μ A -σ A ,μ A +σ A ].

[0025] Furthermore, the shape of the cave is controlled by the rotation matrix R:

[0026]

[0027] The rotation angles α and β are randomly selected from a narrow range [-10°, 10°] to construct a tilted ellipsoid;

[0028] The above parameters (r x ,r y ,r z ,A,α,β) are randomly selected within the current work area, thus performing a large number of earthquake models that are adapted to the geological characteristics of the work area.

[0029] Furthermore, synthetic seismic records are used as training data, and forward cave models are used as labels, 70% of which are set as training sets and 30% as validation sets. CNN neural network is used for training, and the training is completed when the accuracy of the validation set reaches 90%.

[0030] In a second aspect, the present invention provides an artificial intelligence identification system for carbonate caves, comprising:

[0031] The data statistics module is used to obtain the distribution characteristics of the caves by superimposing the post-stack seismic inversion and the instantaneous amplitude at the same level, and to count their size and amplitude;

[0032] The model acquisition module is used to obtain an initial layered model using seismic horizons and post-stack seismic inversion, and to superimpose the initial layered model with the forward cave model to obtain a forward cave model based on the work area;

[0033] The training module is used to convolve the forward-modeled cave model based on the work area with the seismic wavelet of the work area to synthesize the seismic records, use the forward-modeled cave model as a label, and use the CNN neural network to train the seismic records;

[0034] The recognition module is used to finally input the actual seismic data into the trained neural network to obtain a finely described carbonate cave data volume.

[0035] In a third aspect, the present invention provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of an artificial intelligence identification method for carbonate caves when executing the computer program.

[0036] In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a carbonate cave artificial intelligence identification method.

[0037] Compared with the prior art, the present invention has the following technical effects:

[0038] The present invention proposes a method for establishing a carbonate cave forward model based on the geological characteristics of the work area, so that the establishment of the forward model of the cave is more in line with the actual geological conditions; the forward model is used as a training label to participate in the cave prediction under the CNN neural network method, which enhances the accuracy and credibility of cave identification. In actual data application, the location and size of the cave are accurately identified. It provides a research basis for the next step of geological research and oil and gas exploration. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of the present invention.

[0040] Figure 2 It is a flow chart of the artificial intelligence identification method of carbonate caves based on actual geological characteristics.

[0041] Figure 3 This is the forward cave model distribution map.

[0042] Figure 4 It is a cave distribution map based on actual earthquake records and artificial intelligence identification. DETAILED DESCRIPTION

[0043] The specific embodiments of the present invention will be further described in detail below in conjunction with the accompanying drawings. The illustrative examples and descriptions of the present invention are used to explain the present invention, but are not intended to limit the present invention.

[0044] In view of the problem that the current prediction of carbonate caves is easily affected by geological and non-geological information other than cave information, resulting in low recognition accuracy, a fast and accurate artificial intelligence recognition method for carbonate caves is proposed, and the training seismic data is effectively constrained based on actual geological characteristics to identify caves, laying a good foundation for subsequent structural interpretation and reservoir prediction.

[0045] Figure 1 This is a flow chart of the artificial intelligence identification method of carbonate caves based on actual geological characteristics. This method mainly identifies carbonate caves through cave forward simulation and CNN neural network.

[0046] Firstly, the post-stack seismic inversion and instantaneous amplitude are superimposed at the same level to obtain the distribution characteristics of the caves, and their sizes and amplitudes are counted. Under their constraints, the cave model is forward modeled.

[0047] Secondly, the initial layered model is obtained by using seismic horizons and post-stack seismic inversion, which is superimposed on the cave model to obtain a forward model based on the work area.

[0048] Then, the forward model is convolved with the seismic wavelet to synthesize the seismic record, and the forward cave model is used as a label to train the seismic record using the CNN neural network.

[0049] Finally, the actual seismic data is input into the trained neural network to obtain a finely characterized carbonate cave data volume.

[0050] Figure 3 It is a forward model distribution map of the cave. The size and amplitude of the cave must meet the actual geological characteristics of the work area.

[0051] Figure 4It is a karst cave distribution map based on actual seismic records and artificial intelligence identification. Although an obvious "beaded" structure can be seen from the actual seismic records, the size, location, boundaries and other characteristics of the cave are still unclear. The results of artificial intelligence identification can better reflect the distribution characteristics of the cave, and its size, location, boundaries and other characteristics become clear. The identification accuracy of the karst cave reservoir is further improved, laying a good foundation for reservoir prediction.

[0052] The technical invention scheme is mainly implemented through the following steps and methods: Figure 2 ).

[0053] Step 1: On the basis of post-stack seismic inversion, the instantaneous amplitude of seismic records is stacked at the same level to obtain the distribution of carbonate caves;

[0054] Step 2: Count the size of carbonate caves (r x ,r y ,r z ) and the amplitude abnormality A, and the average value (μ x ,μ y ,μ z ,μ A ), standard deviation (σ x ,σ y ,σ z ,σ A ).

[0055]

[0056] Step 3: Under the constraint of seismic horizon, the average value of the amplitude of the interlayer inversion results is calculated and set as the amplitude value of the layer segment to obtain the initial layered seismic model of the layered medium;

[0057] Step 4: On the basis of the initial layered seismic model, a seismic model containing multiple cave characteristics (size and amplitude) is superimposed to obtain a forward cave model based on the geological conditions of the work area.

[0058] The main feature of caves in seismic data is the "beaded" structure, that is, the high amplitude value that is concave downward. In the seismic profile, each cave feature can be approximated as a vertically elongated ellipsoid. Inside the ellipsoid, the seismic waveform extends downward and shows a high amplitude value. Therefore, we can use the ellipsoid calculation formula to forward model carbonate caves:

[0059]

[0060] In order to meet the actual geological characteristics of the work area, the radius of the ellipsoid in three directions should satisfy r x ∈[μ x -σ x ,μ x +σ x ],ry ∈[μ y -σ y ,μ y +σ y ],r z ∈[μ z -σ z ,μ z +σ z ], and the amplitude inside the ellipsoid should satisfy A∈[μ A -σ A ,μ A +σ A ].

[0061] In the actual exploration process, the shape of the cave is not necessarily perpendicular to the seismic phase axis, and some of them are tilted at a small angle, which can be controlled by the rotation matrix R:

[0062]

[0063] The rotation angles α and β are randomly selected from a narrow range [-10°, 10°] to construct a slightly tilted ellipsoid.

[0064] The above parameters (r x ,r y ,r z ,A,α,β) are randomly selected within the current work area, thus performing a large number of earthquake models that adapt to the geological characteristics of the work area, ensuring the accuracy and transferability of this method.

[0065] Step 5: Convolve the forward cave model based on the geological conditions of the work area with the seismic wavelet of the work area to obtain a synthetic seismic record;

[0066] Step 6: Use synthetic seismic records as training data and forward cave models as labels, with 70% set as training set and 30% set as validation set. Use CNN neural network for training. When the accuracy of the validation set reaches 90%, the training is completed.

[0067] Step 7: Use the trained neural network to identify the caves in the work area, and finely characterize their location, boundaries, size and other distribution characteristics to achieve the purpose of carbonate cave identification.

[0068] In yet another embodiment of the present invention, a carbonate cave artificial intelligence identification system is provided, which can be used to implement the above-mentioned carbonate cave artificial intelligence identification method. Specifically, the system includes:

[0069] The data statistics module is used to obtain the distribution characteristics of the caves by superimposing the post-stack seismic inversion and the instantaneous amplitude at the same level, and to count their size and amplitude;

[0070] The model acquisition module is used to obtain an initial layered model using seismic horizons and post-stack seismic inversion, and to superimpose the initial layered model with the forward cave model to obtain a forward cave model based on the work area;

[0071] The training module is used to convolve the forward-modeled cave model based on the work area with the seismic wavelet of the work area to synthesize the seismic records, use the forward-modeled cave model as a label, and use the CNN neural network to train the seismic records;

[0072] The recognition module is used to finally input the actual seismic data into the trained neural network to obtain a finely described carbonate cave data volume.

[0073] The division of modules in the embodiments of the present invention is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional module in each embodiment of the present invention may be integrated into one processor, or may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.

[0074] In another embodiment of the present invention, a computer device is provided, the computer device including a processor and a memory, the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, which are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions in a computer storage medium to implement a corresponding method flow or a corresponding function; the processor described in the embodiment of the present invention can be used for the operation of a carbonate cave artificial intelligence identification method.

[0075] In another embodiment of the present invention, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in a computer device and, of course, an extended storage medium supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the above-mentioned embodiment of a carbonate cave artificial intelligence identification method.

[0076] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0077] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0078] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0079] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A carbonate cave artificial intelligence identification method, characterized in that: include: The distribution characteristics of the caves are obtained by superimposing the post-stack seismic inversion with the instantaneous amplitude at the same level, and their sizes and amplitudes are counted; The initial layered model is obtained by using seismic horizons and post-stack seismic inversion, and the initial layered model is superimposed on the forward cave model to obtain a forward cave model based on the work area; The forward-modeled karst cave model based on the work area is convolved with the seismic wavelet of the work area to synthesize the seismic records. The forward-modeled karst cave model is used as a label to train the seismic records using the CNN neural network. Finally, the actual seismic data is input into the trained neural network to obtain a finely characterized carbonate cave data volume.

2. The artificial intelligence identification method for carbonate caves according to claim 1 is characterized in that: The distribution characteristics of the caves are obtained by superimposing the post-stack seismic inversion and the instantaneous amplitude at the same level, and their size and amplitude are statistically analyzed: Based on the post-stack seismic inversion, the instantaneous amplitude of the seismic records is stacked at the same level to obtain the distribution of carbonate caves; Statistical size of carbonate caves (r x ,r y ,r z ) and the amplitude anomaly size A, and the average value of the cave size and the seismic amplitude anomaly (μ x ,μ y ,μ z ,μ A ), and standard deviation (σ x ,σ y ,σ z ,σ A ); Where μ x ,μ y ,μ z ,μ A are the average values ​​of cave size and earthquake amplitude, σ x , σ y , σ z , σ A are the standard deviations of cave size and earthquake amplitude, are the size of the ith cave and the earthquake amplitude, N is the number of caves, and x, y, and z are the east-west, north-south, and vertical directions, respectively.

3. The artificial intelligence identification method for carbonate caves according to claim 1 is characterized in that: The initial layered model is obtained using seismic horizons and post-stack seismic inversion: Under the constraint of seismic horizon, the average amplitude of the interlayer inversion results is calculated and set as the amplitude value of the layer segment to obtain the initial layered seismic model of the layered medium.

4. The artificial intelligence identification method for carbonate caves according to claim 1 is characterized in that: The initial layered model is superimposed on the forward cave model to obtain a forward cave model based on the work area, and a seismic model containing multiple cave sizes and amplitude characteristics is superimposed.

5. The artificial intelligence identification method for carbonate caves according to claim 4 is characterized in that: The main feature of caves in seismic data is the "beaded" structure, that is, the high amplitude value that is concave downward; in the seismic profile, each cave is characterized by a vertically elongated ellipsoid; inside the ellipsoid, the seismic waveform extends downward and shows a high amplitude value; the carbonate cave is forward modeled using the ellipsoid calculation formula: The spherical coordinates of the ellipsoid are (x, y, z), and the radii of the ellipsoid in three directions satisfy r x ∈[μ x -σ x ,μ x +σ x ],r y ∈[μ y -σ y ,μ y +σ y ],r z ∈[μ z -σ z ,μ z +σ z ], and the amplitude inside the ellipsoid should satisfy A∈[μ A -σ A ,μ A +σ A ].

6. The artificial intelligence identification method for carbonate caves according to claim 4 is characterized in that: The shape of the cave is controlled by the rotation matrix R: The rotation angles α and β are randomly selected from a narrow range [-10°, 10°] to construct a tilted ellipsoid; The above parameters (r x ,r y ,r z ,A,α,β) are randomly selected within the current work area, thus performing a large number of earthquake models that are adapted to the geological characteristics of the work area.

7. The artificial intelligence identification method for carbonate caves according to claim 1 is characterized in that: The synthetic seismic records are used as training data, and the forward cave model is used as the label, 70% of which is set as the training set and 30% is set as the validation set. The CNN neural network is used for training. When the accuracy of the validation set reaches 90%, the training is completed.

8. An artificial intelligence identification system for carbonate caves, characterized in that: include: The data statistics module is used to obtain the distribution characteristics of the caves by superimposing the post-stack seismic inversion and the instantaneous amplitude at the same level, and to count their size and amplitude; The model acquisition module is used to obtain an initial layered model using seismic horizons and post-stack seismic inversion, and to superimpose the initial layered model with the forward cave model to obtain a forward cave model based on the work area; The training module is used to convolve the forward-modeled cave model based on the work area with the seismic wavelet of the work area to synthesize the seismic records, use the forward-modeled cave model as a label, and use the CNN neural network to train the seismic records; The recognition module is used to finally input the actual seismic data into the trained neural network to obtain a finely described carbonate cave data volume.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the carbonate cave artificial intelligence identification method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the carbonate cave artificial intelligence identification method as described in any one of claims 1 to 7 are implemented.

Images