A seismic waveform classification calculation method based on a multi-frequency data body

Abstract

The present application relates to a kind of seismic waveform classification calculation methods based on multi-frequency data body, belong to the high-quality reservoir prediction technical field of energy exploration field such as oil and gas field exploration, shale gas exploration, geothermal exploration, the method is first based on the 3D seismic data of target interval in study area, using time-frequency analysis method calculates 3 common frequency data bodies such as low frequency, medium frequency, high frequency, then sequentially carries out waveform clustering analysis method based on neural network training for 3 common frequency data bodies and carries out the binary analysis of seismic waveform, the study area is divided into 8 sub-regions, then using original data body carries out waveform overall correlation analysis and merging of 8 sub-regions, finally using different color scale represents the value of sub-region, drawing, using image representation waveform classification result.

Description

Technical Field

[0001] This invention relates to the field of oil and gas field exploration technology, particularly to the fields of energy exploration such as oil and gas field exploration, shale gas exploration, and geothermal exploration, and specifically to a seismic waveform classification and calculation method based on multi-frequency data volumes. Background Technology

[0002] Seismic facies are the comprehensive response characteristics of sedimentary facies in seismic data. Seismic facies technology based on waveform classification classifies the waveform shape of seismic traces in a target layer, and uses this classification to characterize sedimentary facies zones over a large area or lithofacies zones within a small area.

[0003] Seismic waveform classification is based on the relationship between parameters such as frequency, amplitude, and phase of seismic signals and waveform characteristics. It combines techniques such as neural networks to process seismic waveform information to determine the spatial similarity of seismic waveforms. Conventional seismic waveform classification algorithms often employ waveform clustering analysis methods based on neural network training or waveform classification methods based on vector space models, such as neural networks, logistic regression, and support vector machines.

[0004] There are many methods for classifying seismic waveforms, mainly divided into unsupervised classification and supervised classification. The method in this invention patent belongs to the unsupervised classification method. When prior information is incomplete or nonexistent, classification through unsupervised learning is one of the most effective methods. Two common unsupervised learning methods are: ① Self-organizing neural network clustering, which uses unsupervised learning to train and generate a low-dimensional (usually two-dimensional), discrete training sample input space, called a map, realizing an ordered mapping from high-dimensional distribution to a regular low-dimensional grid, which is a dimensionality reduction method; ② Gaussian mixture model, which is a parametric model of probability distribution, provides greater flexibility and accuracy in modeling than traditional unsupervised clustering algorithms. Unsupervised classification algorithms are mainly applied to seismic facies analysis of post-stack seismic data because post-stack seismic data is simple and computationally inexpensive, and these algorithms can quickly obtain the necessary results.

[0005] In summary, current research on seismic waveform classification mainly focuses on in-depth research on classification algorithms, while neglecting in-depth mining of the input data volume. Summary of the Invention

[0006] This invention relates to a seismic waveform classification and calculation method based on multi-frequency data volumes, belonging to the field of high-quality reservoir prediction technology in energy exploration, such as oil and gas field exploration, shale gas exploration, and geothermal exploration. The method first uses three-dimensional seismic data of the target layer in the study area and employs time-frequency analysis to obtain three common-frequency data volumes: low-frequency, mid-frequency, and high-frequency. Then, waveform binary analysis based on a neural network-trained waveform clustering method is sequentially performed on these three common-frequency data volumes. Specifically, a single waveform clustering analysis divides the waveforms of the target area into two categories, dividing the study area into eight sub-regions. Next, the original data volumes are used to perform overall correlation analysis and merging of the waveforms in the eight sub-regions. Finally, different color codes are used to represent the values ​​of the sub-regions, and the results are plotted using an image representation.

[0007] The specific steps of this invention include:

[0008] (1) Input the three-dimensional seismic data of the target layer in the study area, perform time-frequency analysis on the three-dimensional data volume, and obtain three common frequency data volumes corresponding to the original data volume based on the dominant frequency, which are approximately low-frequency data volume, medium-frequency data volume and high-frequency data volume;

[0009] (2) For low-frequency data volumes, first carry out denoising processing based on the three-dimensional edge-preserving denoising method, and then use the waveform clustering analysis method based on neural network training to carry out waveform clustering analysis and perform binary analysis of the waveform in the whole region. That is, by dividing the waveform into two types, the study area is divided into two sub-regions.

[0010] (3) For the medium frequency data volume, the waveform clustering analysis method based on neural network training is used for the two sub-regions in step (2) to carry out waveform clustering analysis and perform bisection analysis of the waveforms in the sub-regions to obtain 4 sub-regions. That is, after this step, the study area is divided into 4 sub-regions.

[0011] (4) For high-frequency data volumes, waveform clustering analysis based on neural network training is used for the four sub-regions in step (3) to carry out waveform clustering analysis and perform binary analysis of the waveforms in the sub-regions to obtain eight sub-regions. That is, after this step, the study area is divided into eight sub-regions.

[0012] (5) Using the original data volume of the 8 sub-regions divided in step (4) (without frequency division processing), perform correlation analysis and merging of the 8 sub-regions;

[0013] (6) Use different color marks to represent the values ​​of sub-regions, draw a diagram, and use images to represent the final waveform classification results.

[0014] A waveform classification calculation method based on multi-frequency data volumes has the following characteristics, mainly manifested as follows:

[0015] (1) For the first time in the field of waveform classification, this invention uses three common frequency data volumes, which approximately represent the low frequency band data volume, the mid frequency band data volume, and the high frequency band data volume, and the original data volume (without frequency division processing) to perform waveform classification calculation.

[0016] (2) This invention makes full use of the advantages of low-frequency data volume representing macroscopic features and high-frequency data volume representing detailed features. In the waveform classification process design, the macroscopic (overall partitioning) is performed first, followed by the microscopic (local partitioning), and finally the original data volume (respecting the waveform features of the original data volume) is used to carry out sub-region merging analysis. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the technical process of the present invention.

[0018] Figure 2 This is a waveform classification planar diagram of the second section of the Qixia Formation in a certain area of ​​Sichuan Province, calculated using the method of this invention. Detailed Implementation

[0019] Example 1

[0020] A seismic waveform classification calculation method based on multi-frequency data volumes, comprising the following steps:

[0021] Step 1: Input the three-dimensional seismic data of the target segment in the study area, denoted as array A(x,y,t);

[0022] Step 2: Calculate the dominant frequency (Fmain) of array A(x,y,t) using Fourier transform analysis.

[0023] Step 3: Perform time-frequency analysis on array A(x,y,t) using the time-frequency wavelet transform method to obtain the time-frequency analysis array AW(x,y,t,f);

[0024] Step 4: Extract the common frequency data volumes of three different frequencies.

[0025] Step 4-1: Denote the frequency of the first common-frequency data block as f1, the frequency of the second common-frequency data block as f2, and the frequency of the third common-frequency data block as f3.

[0026] f1=int(Fmain- Fmain / 3), f2=int(Fmain), f3=int(Fmain+Fmain / 3)

[0027] Step 4-2: Extract three common frequency data volumes from the time-frequency analysis array AW(x,y,t,f) in Step 3: AW(x,y,t,f1), AW(x,y,t,f2), and AW(x,y,t,f3). Since f1, f2, and f3 are fixed values, the arrays AW(x,y,t,f1), AW(x,y,t,f2), and AW(x,y,t,f3) are 3-dimensional arrays. Let AW_low(x,y,t) = AW(x,y,t,f1), AW_med(x,y,t) = AW(x,y,t,f2), and AW_hig(x,y,t) = AW(x,y,t,f3).

[0028] Step 5: Perform a bisection analysis of the full-area waveform using the array AW_low(x,y,t).

[0029] Step 5-1: Apply a three-dimensional edge-preserving denoising method to the array AW_low(x,y,t). The denoised value is denoised as AW_low_n(x,y,t).

[0030] Step 5-2: For the array AW_low_n(x,y,t), in the time t direction, a waveform clustering analysis method based on neural network training is used to perform waveform clustering analysis. The number of waveform categories is set to 2, and the classification result is recorded as Wave1(x,y). The array Wave1(x,y) uses 2 values, 1 or 2, to represent 2 types of waveforms.

[0031] Step 6: Use the array AW_med(x,y,t) to perform binary analysis of the waveform in different regions.

[0032] Step 6-1: Denoise the array AW_med(x,y,t) using a three-dimensional edge-preserving denoising method. The denoised value is denoised as AW_med_n(x,y,t).

[0033] Step 6-2: Extract the region where Wave1(x,y)=1. For this region, use the array AW_med_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_1(x,y). The array Wave1_1 uses two values, 11 or 12, representing the two types of waveforms.

[0034] Step 6-3: Extract the region where Wave1(x,y)=2. For this region, use the array AW_med_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_2(x,y). The array Wave1_2 uses two values, 21 or 22, to represent the two types of waveforms.

[0035] Step 7: Use the array AW_hig(x,y,t) to perform binary analysis of the regional waveforms.

[0036] Step 7-1: The array AW_hig(x,y,t) is denoised using a three-dimensional edge-preserving denoising method. The denoised value is denoted as AW_hig_n(x,y,t).

[0037] Step 7-2: Extract the region where Wave1_1(x,y)=11. For this region, use the array AW_hig_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_1_1(x,y). The array Wave1_1_1 uses two values, 111 or 112, representing the two types of waveforms.

[0038] Step 7-3: Extract the region where Wave1_1(x,y)=12. For this region, use the array AW_hig_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_1_2(x,y). The array Wave1_1_2 uses two values, 121 or 122, representing the two types of waveforms.

[0039] Step 7-4: Extract the region where Wave1_2(x,y)=21. For this region, use the array AW_hig_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_2_1(x,y). The array Wave1_2_1 uses two values, 211 or 212, representing the two types of waveforms.

[0040] Step 7-5: Extract the region where Wave1_2(x,y)=22. For this region, use the array AW_hig_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_2_2(x,y). The array Wave1_2_2 uses two values, 221 or 222, to represent the two types of waveforms.

[0041] Step 8: Merge the 8 regions obtained in Step 7, denoted as Wave0(x,y). The array Wave0(x,y) has 8 values: 111, 112, 121, 122, 211, 212, 221, 222, which correspond to the 8 regions in Step 7 respectively.

[0042] Step 9: Based on A(x,y,t) from Step 1 and Wave0(x,y) from Step 8, divide A(x,y,t) into 8 sub-regions. Calculate the correlation coefficient between the 8 sub-regions using covariance matrix analysis. Merge sub-regions with correlation coefficients greater than 0.9 and update the value of Wave0(x,y). For example, if the overall correlation coefficient between the waveforms of the first and second sub-regions is greater than 0.9, modify the value of Wave0(x,y) of the second sub-region to match the value of Wave0(x,y) of the first sub-region, thus merging the first and second sub-regions.

[0043] Step 10: Output the Wave0(x,y) array, and use different color codes to represent the values ​​of the Wave0(x,y) array. Plot the graph and use the image to represent the final waveform classification result.

[0044] Example 2

[0045] Figure 2 This is a waveform classification plan of the second member of the Qixia Formation in a research area of ​​Sichuan. The second, third, and fourth categories of regions in the figure well represent the development areas of high-quality oil and gas reservoirs.

[0046] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. The specific steps of a seismic waveform classification calculation method based on multi-frequency data volumes include: Step 1: Input the three-dimensional seismic data of the target segment in the study area, denoted as array A(x,y,t); Step 2: Calculate the dominant frequency (Fmain) of array A(x,y,t) using Fourier transform analysis. Step 3: Perform time-frequency analysis on array A(x,y,t) using the time-frequency wavelet transform method to obtain the time-frequency analysis array AW(x,y,t,f); Step 4: Extract the common frequency data volumes of three different frequencies. Step 4-1: Denote the frequency of the first common-frequency data block as f1, the frequency of the second common-frequency data block as f2, and the frequency of the third common-frequency data block as f3. f1=int(Fmain- Fmain / 3), f2=int(Fmain), f3=int(Fmain+Fmain / 3) Step 4-2: Extract three common frequency data volumes from the time-frequency analysis array AW(x,y,t,f) in Step 3: AW(x,y,t,f1), AW(x,y,t,f2), and AW(x,y,t,f3). Since f1, f2, and f3 are fixed values, the arrays AW(x,y,t,f1), AW(x,y,t,f2), and AW(x,y,t,f3) are 3-dimensional arrays. Let AW_low(x,y,t) = AW(x,y,t,f1), AW_med(x,y,t) = AW(x,y,t,f2), and AW_hig(x,y,t) = AW(x,y,t,f3). Step 5: Perform a bisection analysis of the full-area waveform using the array AW_low(x,y,t). Step 5-1: Apply a three-dimensional edge-preserving denoising method to the array AW_low(x,y,t). The denoised value is denoised as AW_low_n(x,y,t). Step 5-2: For the array AW_low_n(x,y,t), in the time t direction, a waveform clustering analysis method based on neural network training is used to perform waveform clustering analysis. The number of waveform categories is set to 2, and the classification result is recorded as Wave1(x,y). The array Wave1(x,y) uses 2 values, 1 or 2, to represent 2 types of waveforms. Step 6: Use the array AW_med(x,y,t) to perform binary analysis of the waveform in different regions. Step 6-1: Denoise the array AW_med(x,y,t) using a three-dimensional edge-preserving denoising method. The denoised value is denoised as AW_med_n(x,y,t). Step 6-2: Extract the region where Wave1(x,y)=1. For this region, use the array AW_med_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_1(x,y). The array Wave1_1 uses two values, 11 or 12, representing the two types of waveforms. Step 6-3: Extract the region where Wave1(x,y)=2. For this region, use the array AW_med_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_2(x,y). The array Wave1_2 uses two values, 21 or 22, to represent the two types of waveforms. Step 7: Use the array AW_hig(x,y,t) to perform binary analysis of the regional waveforms. Step 7-1: The array AW_hig(x,y,t) is denoised using a three-dimensional edge-preserving denoising method. The denoised value is denoted as AW_hig_n(x,y,t). Step 7-2: Extract the region where Wave1_1(x,y)=11. For this region, use the array AW_hig_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_1_1(x,y). The array Wave1_1_1 uses two values, 111 or 112, representing the two types of waveforms. Step 7-3: Extract the region where Wave1_1(x,y)=12. For this region, use the array AW_hig_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_1_2(x,y). The array Wave1_1_2 uses two values, 121 or 122, representing the two types of waveforms. Step 7-4: Extract the region where Wave1_2(x,y)=21. For this region, use the array AW_hig_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_2_1(x,y). The array Wave1_2_1 uses two values, 211 or 212, representing the two types of waveforms. Step 7-5: Extract the region where Wave1_2(x,y)=22. For this region, use the array AW_hig_n(x,y,t) to perform waveform clustering analysis in the time t direction using a waveform clustering analysis method based on neural network training. Set the number of waveform categories to 2. Record the classification results as Wave1_2_2(x,y). The array Wave1_2_2 uses two values, 221 or 222, to represent the two types of waveforms. Step 8: Merge the 8 regions obtained in Step 7, denoted as Wave0(x,y). The array Wave0(x,y) has 8 values: 111, 112, 121, 122, 211, 212, 221, 222, which correspond to the 8 regions in Step 7 respectively. Step 9: Based on A(x,y,t) from Step 1 and Wave0(x,y) from Step 8, divide A(x,y,t) into 8 sub-regions, calculate the correlation coefficient between the 8 sub-regions using the covariance matrix analysis method, merge the sub-regions with a correlation coefficient greater than 0.9, and update the value of Wave0(x,y). Step 10: Output the Wave0(x,y) array, and use different color codes to represent the values ​​of the Wave0(x,y) array. Plot the graph and use the image to represent the final waveform classification result.

Images