Intelligent diffraction wave separation method based on common imaging point gathers

Abstract

The present invention discloses an intelligent diffraction wave separation method based on common imaging point gathers, which relates to the field of exploration geophysical technology, and includes the following steps: obtaining input data; calculating full wave field and reflection wave seismic records using the input data; calculating full wave field common imaging point gathers, full wave field imaging results and reflection wave common imaging point gathers; constructing a 4-layer convolutional neural network structure; iteratively training the convolutional neural network parameters using the generated training data set; applying the trained neural network on the test data set to predict the reflection wave characteristics presenting a "smiley face" shape; obtaining diffraction wave common imaging point gathers using an adaptive subtraction method, and superimposing the gathers to obtain a diffraction wave imaging profile. The present invention can obtain diffraction wave dip angle domain common imaging point gathers with a high signal-to-noise ratio and a high-precision underground small-scale geological body imaging profile, and improves the imaging resolution of the diffraction target body while ensuring the calculation efficiency.

Description

Technical Field

[0001] The invention relates to the field of exploration geophysical technology, and in particular to an intelligent diffraction wave separation method based on common imaging point gathers. Background Art

[0002] Reflection waves and diffraction waves are two typical seismic responses in seismic exploration. Reflection waves originate from smooth and continuous interfaces underground and are the main targets of current seismic inversion and imaging; diffraction waves originate from small-scale discontinuous geological structures and irregular geological bodies (faults, cracks, fractures, etc.) underground and carry high-resolution seismic information. With the needs of oil and gas field exploration and development, conventional exploration methods that use reflection waves as the main research object can no longer meet the accuracy requirements, while diffraction wave imaging has the ability to detect small-scale geological bodies. Therefore, making full use of the diffraction wave response in seismic data and realizing separate imaging of diffraction waves is of great significance for improving the seismic imaging resolution of small-scale discontinuous structures and high-precision seismic interpretation under complex structural conditions. However, the diffraction wave energy in seismic data is often submerged in the relatively strong reflection information.

[0003] In the seismic data processing process, weak amplitude diffraction waves carrying underground high-resolution information are often regarded as noise and filtered out or suppressed, resulting in low resolution imaging of small-scale discontinuous and inhomogeneous geological bodies in the underground. The common solution is to extract and separate the reflection wave and diffraction wave based on the characteristic differences in the pre-stack seismic gathers and migration process. The method of extracting diffraction waves in pre-stack gathers has the problem of requiring high signal-to-noise ratio original seismic data and low computational efficiency. In the migration process, diffraction wave imaging can be achieved by improving the migration operator, but conventional wave equation migration has the disadvantages of geometric diffusion loss of propagation energy, insufficient illumination of deep areas, and the need for high coverage times of seismic data. Various types of gathers can also be extracted from the migration process, and diffraction wave extraction can be achieved by identifying the energy near the stable phase vertex of the reflection wave and suppressing the reflection wave, but the effect of this type of wave field separation method depends largely on the reflection wave cut-off radius. The diffraction wave response from the underground small-scale discontinuous geological body is difficult to effectively extract and separate from the seismic data, which greatly reduces the imaging accuracy and resolution of the diffraction target.

[0004] Therefore, it is urgent to study a diffraction wave separation and extraction method to solve the problem that the weaker energy diffraction waves in the seismic records are covered by the strong reflection response, making it difficult to achieve diffraction wave separation and high-precision imaging of small-scale geological bodies, so as to achieve high-resolution seismic imaging of small-scale structures such as underground faults and pinch-outs by diffraction waves, so as to meet the current detailed exploration of structural and lithological oil and gas reservoirs. Summary of the invention

[0005] In order to solve the above technical problems, the present invention discloses an intelligent diffraction wave separation method based on common imaging point gathers. The present invention first calculates common imaging point gathers containing only reflection waves and both reflection waves and scattered waves based on a preset scattering point velocity model through full wavefield forward simulation and Gaussian beam common imaging point gather offset, then uses the reflection wave common imaging point gather as the input of the deep neural network, and uses the imaging gather containing both reflection waves and scattered waves as the output of the intelligent network, uses a synthetic training data set to train the neural network structure, uses the obvious geometric feature differences of the reflection wave and diffraction wave responses on the common imaging point gathers, extracts the reflection wave features with a "smiley face" shape with a stable phase vertex, and finally uses adaptive subtraction to calculate the diffraction wave common imaging point gathers and imaging profiles, and finally obtains high-resolution imaging results of the diffraction target.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] According to a first aspect of an embodiment of the present invention, an intelligent diffraction wave separation method based on common imaging point gathers is proposed.

[0008] In one embodiment, an intelligent diffraction wave separation method based on common imaging point gathers includes the following steps:

[0009] S1: Get input data;

[0010] S2: Calculate full wavefield and reflection wave seismic records using input data;

[0011] S3: Calculate the full wavefield common imaging point gathers, full wavefield imaging results and reflection wave common imaging point gathers;

[0012] S4: Construct a 4-layer convolutional neural network structure;

[0013] S5: Iteratively train the convolutional neural network parameters constructed in step S4 using the training data set generated in step S3;

[0014] S6: Apply the neural network trained in step S5 to the test data set to identify the reflection wave features of the "smiley face" shape;

[0015] S7: Use the adaptive subtraction method to obtain the diffraction wave common imaging point gathers, and superimpose the gathers to obtain the diffraction wave imaging profile.

[0016] Optionally, in step S1, the input data includes: velocity models including and excluding scattering points, smooth migration velocity models, source functions and seismic wavelets.

[0017] Optionally, in step S2, the calculating of the full wave field and the reflected wave seismic record by using the input data includes using velocity models including and excluding scattering points, source functions and seismic wavelets to calculate the full wave field and the reflected wave seismic record by solving an acoustic medium wave equation, wherein the acoustic wave wave equation is:

[0018]

[0019] Among them, v is the longitudinal wave propagation velocity, p is the pressure, and f is the source function. represents the time second derivative, is the Laplace operator.

[0020] Optionally, in step S3, full wavefield common imaging point gathers and full wavefield imaging results are calculated by a migration algorithm using full wavefield seismic records and a smoothed velocity model, and reflection wave common imaging point gathers are calculated by a migration algorithm using reflection wave seismic records and a smoothed velocity model, and the migration algorithm is:

[0021]

[0022] Among them, I(x; x s ; θ) is the imaging result, X r represents the detection point position, x s is the source position, x is the imaging point position, θ is the incident angle, p(x r ;x s ; t) is the pressure record, τ is the round trip time, W(x r ;x;x s ; θ) is the migration imaging weighting coefficient.

[0023] Optionally, in step S5, during iterative training, the full wavefield dip domain common imaging point gathers containing reflection wave and diffraction wave responses are used as neural network input, and the dip domain common imaging point gathers containing only reflection wave responses are used as neural network output.

[0024] Optionally, the adaptive subtraction method is to adaptively subtract the reflection wave gathers predicted by the neural network from the full wavefield dip domain gathers.

[0025] According to a second aspect of the embodiments of the present invention, an electronic device is provided.

[0026] In one embodiment, the electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor implements the steps of any of the above methods when executing the program.

[0027] According to a third aspect of an embodiment of the present invention, a computer-readable storage medium is provided.

[0028] In one embodiment, the computer-readable storage medium stores a program for executing any of the above methods.

[0029] The beneficial effects of the present invention are:

[0030] 1. The present invention can successfully separate the diffraction waves from underground small-scale geological structures and submerged by the reflected wave energy, and finally obtain high-precision diffraction body imaging results.

[0031] 2. This patent applies convolutional neural network to separate diffraction waves based on the geometric feature differences of diffraction wave and reflection wave responses in the dip common imaging point gathers, abandoning the traditional method of analyzing the differences in kinematic and dynamic characteristics of reflection waves and diffraction waves in different seismic gathers in order to ensure the success of diffraction wave separation.

[0032] 3. The present invention firstly calculates the full wave field and the reflected wave dip domain common imaging point gathers using the velocity model with and without artificial scattering points, and then uses the two gathers as the input and output of the neural network to train the neural network parameters, so that the network can successfully identify the reflection wave characteristics with a stable phase vertex in the shape of a "smiley face" during the training process, and finally the trained network is applied to the test area. Compared with the traditional diffraction wave separation method based on vertex removal, the present invention can identify the waveform characteristics of a larger lateral range near the stable phase vertex of the reflected wave, and the reflection wave feature recognition radius can well adapt to the complex underground structure. Although the neural network takes several hours in the training process, the neural network that captures the geometric characteristics of the reflected wave only takes a few seconds in the application process.

[0033] 4. The present invention further applies an adaptive subtraction method to calculate diffraction wave gathers and imaging results, and finally obtains diffraction wave inclination domain common imaging point gathers with high signal-to-noise ratio and high-precision underground small-scale geological body imaging profiles, thereby improving the imaging resolution of the diffraction target body while ensuring the calculation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A flowchart of an intelligent diffraction wave separation method based on common imaging point gathers is shown in one embodiment of the present invention;

[0035] Figure 2 An end-to-end convolutional neural network structure for diffraction wave separation shown in one embodiment of the present invention;

[0036] Figure 3 A velocity model including scattering points is shown as an embodiment of the present invention;

[0037] Figure 4 A velocity model not including scattering points is shown in an embodiment of the present invention;

[0038] Figure 5A migration velocity model is shown in one embodiment of the present invention;

[0039] Figure 6 The full wavefield common imaging point gather and the reflected wave common imaging point gather are shown in one embodiment of the present invention;

[0040] Figure 7 The predicted reflection wave gather and the adaptive subtraction gather shown in one embodiment of the present invention;

[0041] Figure 8 This is a full wavefield imaging result diagram shown in an embodiment of the present invention;

[0042] Fig. 9 This is a diagram showing the reflection wave imaging result according to an embodiment of the present invention;

[0043] Fig.10 This is a diagram showing the scattered wave imaging result according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] Example 1

[0046] Figure 1 The flow of an intelligent diffraction wave separation method based on common imaging point gathers is shown.

[0047] In this embodiment, the intelligent diffraction wave separation method based on common imaging point gathers includes the following steps:

[0048] S1: Get input data;

[0049] S2: Calculate full wavefield and reflection wave seismic records using input data;

[0050] S3: Calculate the full wavefield common imaging point gathers, full wavefield imaging results and reflection wave common imaging point gathers;

[0051] S4: Construct convolutional neural network structure;

[0052] S5: Iteratively train the convolutional neural network parameters constructed in step S4 using the training data set generated in step S3;

[0053] S6: Apply the neural network trained in step S5 to the test data set to identify the reflection wave features of the "smiley face" shape;

[0054] S7: Use the adaptive subtraction method to obtain the diffraction wave common imaging point gathers, and superimpose the gathers to obtain the diffraction wave imaging profile.

[0055] Optionally, in step S1, the input data includes: velocity models including and excluding scattering points, smooth migration velocity models, source functions and seismic wavelets. Figure 3 The velocity model including scattering points is shown. Figure 4 shows the velocity model without scattering points, Figure 5 The migration velocity model is shown.

[0056] Optionally, in step S2, the calculating of the full wave field and the reflected wave seismic record by using the input data includes using velocity models including and excluding scattering points, source functions and seismic wavelets to calculate the full wave field and the reflected wave seismic record by solving an acoustic medium wave equation, wherein the acoustic wave wave equation is:

[0057]

[0058] Among them, v is the longitudinal wave propagation velocity, p is the pressure, and f is the source function. represents the time second derivative, is the Laplace operator.

[0059] Figure 6 The full wavefield common imaging point gather and the reflected wave common imaging point gather are shown.

[0060] Optionally, in step S3, full wavefield common imaging point gathers and full wavefield imaging results are calculated by a migration algorithm using full wavefield seismic records and a smoothed velocity model, and reflection wave common imaging point gathers are calculated by a migration algorithm using reflection wave seismic records and a smoothed velocity model, and the migration algorithm is:

[0061]

[0062] Among them, I(x; x s ; θ) is the imaging result, X r represents the detection point position, x s is the source position, x is the imaging point position, θ is the incident angle, p(x r ;x s ; t) is the pressure record, τ is the round trip time, W(x r ;x;x s ; θ) is the migration imaging weighting coefficient.

[0063] Figure 2 An end-to-end convolutional neural network architecture for diffraction wave separation is shown.

[0064] In this embodiment, in step S4, the convolutional neural network structure has 4 layers.

[0065] Optionally, in step S5, during iterative training, the full wavefield dip domain common imaging point gathers containing reflection wave and diffraction wave responses are used as neural network input, and the dip domain common imaging point gathers containing only reflection wave responses are used as neural network output.

[0066] Figure 7 The reflected wave dip domain gathers predicted by the neural network and the diffraction wave dip domain gathers obtained by adaptively subtracting the network predicted reflection wave gathers from the full wavefield gathers are shown. It can be seen that the reflected wave response with a stable phase vertex in the shape of a "smiley face" is clearly identified by the convolutional neural network, and the reflected wave energy in the gathers after adaptive subtraction is suppressed to the greatest extent, while the diffraction wave energy is well preserved.

[0067] Figure 8 The full wavefield imaging result is shown in FIG. Fig. 9 The reflection wave imaging result is shown in FIG. Fig.10 The scattered wave imaging result is shown.

[0068] In this embodiment, the adaptive subtraction method is to adaptively subtract the reflection wave gathers predicted by the neural network from the full wave field dip domain gathers. Fig.10 It can be seen that the reflected wave energy from the underground continuous strata is suppressed, and the diffraction waves with weaker amplitudes from the underground discontinuous small-scale are retained, ultimately obtaining high-precision imaging results of the underground small-scale discontinuous geological bodies.

[0069] Example 2

[0070] This embodiment discloses an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, Figure 1 steps.

[0071] Example 3

[0072] This embodiment discloses a computer-readable storage medium, wherein the computer-readable storage medium stores an executable Figure 1 Procedure with steps shown.

[0073] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or 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 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 codes.

[0074] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0075] Of course, the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by technicians in this technical field within the essential scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. An intelligent diffraction wave separation method based on common imaging point gathers, characterized in that: The following steps are involved: S1: Get input data; S2: Calculate full wavefield and reflection wave seismic records using input data; S3: Calculate the full wavefield common imaging point gathers, full wavefield imaging results and reflection wave common imaging point gathers; S4: Construct a 4-layer convolutional neural network structure; S5: Iteratively train the convolutional neural network parameters constructed in step S4 using the training data set generated in step S3; S6: Apply the neural network trained in step S5 to the test data set to identify the reflection wave features of the "smiley face" shape; S7: Use the adaptive subtraction method to obtain the diffraction wave common imaging point gathers, and superimpose the gathers to obtain the diffraction wave imaging profile; In step S2, the calculation of the full wave field and the reflected wave seismic record using the input data includes using velocity models including and excluding scattering points, source functions and seismic wavelets to calculate the full wave field and the reflected wave seismic record by solving the acoustic medium wave equation, wherein the acoustic medium wave equation is: Among them, v is the longitudinal wave propagation velocity, p is the pressure, and f is the source function. represents the time second derivative, is the Laplace operator; In step S3, full wavefield common imaging point gathers and full wavefield imaging results are calculated by a migration algorithm using full wavefield seismic records and a smoothed velocity model, and reflection wave common imaging point gathers are calculated by a migration algorithm using reflection wave seismic records and a smoothed velocity model. The migration algorithm is: Among them, I(x; x s ; θ) is the imaging result, X r represents the detection point position, x s is the source position, x is the imaging point position, θ is the incident angle, p(x r ;x s ; t) is the pressure record, τ is the round trip time, W(x r ;x;x s ; θ) is the migration imaging weighting coefficient.

2. The intelligent diffraction wave separation method based on common imaging point gathers according to claim 1, characterized in that: In step S1, the input data includes: velocity models including and excluding scattering points, smooth migration velocity models, source functions and seismic wavelets.

3. The intelligent diffraction wave separation method based on common imaging point gathers according to claim 1, characterized in that: In step S5, during iterative training, the full wave field dip domain common imaging point gathers containing reflection wave and diffraction wave responses are used as neural network input, and the dip domain common imaging point gathers containing only reflection wave responses are used as neural network output.

4. The intelligent diffraction wave separation method based on common imaging point gathers according to claim 1, characterized in that: The adaptive subtraction method is to adaptively subtract the reflection wave gathers predicted by the neural network from the full wave field dip domain gathers.

5. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method according to any one of claims 1 to 4 are implemented.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program for executing the method according to any one of claims 1 to 4.

Images