Method for removing surface waves in single-shot seismic data

Abstract

The invention relates to a method for removing surface waves in single-shot seismic data, and belongs to the technical field of seismic exploration. According to the method for removing the surface waves in the single-shot seismic data, the frequency band range of the surface waves is directly investigated on the single shot, and then the seismic data is decomposed according to the octave according to wavelet transform, so that the noise characteristics of the surface waves in the single octave are more prominent and easier to recognize; the seismic data of different frequency bands are suppressed and denoised step by step by adopting different surface wave suppression methods, so that surface wave attenuation can be more thorough. Besides, surface wave attenuation is carried out on the seismic data of the specific frequency bands (the first frequency band and the second frequency band), other frequency bands are not affected, and better amplitude preservation fidelity is achieved.

Description

Technical Field

[0001] The invention relates to a method for removing surface waves in single-shot seismic data, belonging to the technical field of seismic exploration. Background Art

[0002] In the actual seismic exploration process, due to the influence of various factors, the collected seismic data will contain a lot of noise. Among them, the surface wave is a regular interference wave with strong energy. Its wave field is fan-shaped at the near-track of single-shot seismic recording (the near-track refers to the surface wave of the near-track seismic wave), with the characteristics of low apparent velocity, low frequency and strong energy. The existence of surface waves reduces the signal-to-noise ratio and resolution of seismic data.

[0003] Due to the absorption, attenuation and dispersion effects of the medium, the characteristics of the surface wave field vary with time and space, its energy decays rapidly with increasing depth and propagation distance, and the frequency changes relatively from high frequency to low frequency with increasing depth and propagation distance, showing a dispersion phenomenon; the apparent velocity changes regularly with changes in depth and propagation distance, and its event axis is not a cluster of straight lines, but a cluster of curves.

[0004] The methods for suppressing surface rolls based on the frequency difference between surface rolls and effective waves mainly include low-cut frequency, high-pass filtering, band-pass filtering, etc.; the methods for suppressing surface rolls based on the apparent velocity difference include fk filtering, τ-p transformation, etc.; curvelet transformation method, fk domain surface roll suppression method based on empirical mode decomposition (EMD), etc. Most of these methods focus on considering certain characteristics of surface rolls, such as frequency, apparent velocity, etc. With the continuous advancement of seismic exploration technology and changes in production needs, higher requirements are also put forward for the processing effect of seismic data.

[0005] Chinese patent document CN101334483A discloses a method for attenuating Rayleigh wave scattering noise in seismic data processing. This method only performs scattering noise attenuation processing in the frequency band that mainly contains the scattered wave field, and does not affect the original wave field characteristics of other frequency bands. It is a relatively fidelity denoising method. This method comprehensively utilizes linear wave field transformation, wavelet transformation and a-trimmed filtering technology to extract the Rayleigh wave scattered incident wave field and calculate the sub-wave, and then achieves the best estimation of the scattered wave field through the scattered wave field inversion technology, and then subtracts the estimated scattering noise from the original wave field, thereby effectively improving the signal-to-noise ratio of the seismic data. However, this method has the problem of poor amplitude preservation and fidelity. Summary of the invention

[0006] The object of the present invention is to provide a method for removing surface roll in single-shot seismic data, which can solve the problem of poor amplitude preservation and fidelity when removing surface roll in seismic data.

[0007] In order to achieve the above object, the technical solution adopted by the method for removing surface waves in single shot seismic data of the present invention is:

[0008] A method for removing surface waves in single-shot seismic data comprises the following steps:

[0009] (1) Performing static correction and spherical diffusion compensation on the single-shot seismic data to obtain processed single-shot seismic data;

[0010] (2) investigating the frequency band range of the surface wave in the processed single-shot seismic data by octave frequency scanning, and determining the main frequency band of the surface wave; the main frequency band of the surface wave is 0 to x Hz, where x is equal to the sum of the maximum value of the initial frequency band of the surface wave and y, and y is 1 to 2 Hz;

[0011] (3) using discrete wavelet transform to divide the processed single-shot seismic data of the full frequency band according to octaves, the divided frequency bands include a first frequency band and a second frequency band, the first frequency band is 0 to x Hz, and the second frequency band is x to 2×(x+2) Hz;

[0012] (4) Anti-aliasing processing is performed on the seismic data in the first frequency band, and τ-p transformation is used to attenuate the seismic data in the second frequency band.

[0013] The method for removing surface waves from single-shot seismic data of the present invention directly investigates the frequency band range of surface waves on a single shot, and then decomposes the seismic data by octave according to wavelet transform, so that the noise characteristics of surface waves in a single octave are more prominent and easier to identify, and the surface waves can be attenuated more thoroughly by using different surface wave suppression methods to suppress and denoise the seismic data of different frequency bands step by step. In addition, surface wave attenuation is performed on seismic data of specific frequency bands (first frequency band and second frequency band), which does not affect other frequency bands and has better amplitude preservation and fidelity.

[0014] According to a large number of experimental results, in order to improve the denoising rate and fidelity, preferably, the y is 2 Hz.

[0015] In the present invention, the surface wave is seismic data with a velocity of 300 to 900 m / s on a single shot seismic record.

[0016] Preferably, the minimum frequency in the initial frequency band of the surface wave is equal to the minimum value in the frequency range corresponding to the set of data points with a velocity of 300 to 900 m / s in the processed single-shot seismic data, and the maximum frequency in the initial frequency band of the surface wave is equal to the maximum value in the frequency range corresponding to the set of data points with a velocity of 300 to 900 m / s in the processed single-shot seismic data.

[0017] Preferably, the initial frequency band of the surface wave is measured from a processed single-shot seismic record.

[0018] It can be understood that the initial frequency band of the surface wave is determined by a method comprising the following steps: determining a set of data points with a velocity of 300 to 900 m / s in the processed single-shot seismic data, and then determining the frequency range corresponding to the set of data points, the minimum value in the frequency range is the minimum frequency in the initial frequency band of the surface wave, and the maximum value in the frequency range is the maximum frequency in the initial frequency band of the surface wave.

[0019] In order to improve the computational efficiency, continuous wavelets are generally discretized. The core idea is to discretize the translation and scale transformation parameters of the wavelet transform. The transformation process is calculated using two variables i and k of the discrete scaling factor a and the translation factor d. The formula is as follows:

[0020]

[0021] Among them, i∈Z, and a0≠1, since k≠0 and can take any value, in the processing process, a0>1 is taken to obtain the discrete wavelet function ψ i,k Mathematical formula for (t):

[0022]

[0023] From the above formula, the wavelet coefficients after discretization can be deduced as:

[0024]

[0025] Anti-aliasing processing is to establish a model with given speed and frequency to process the noise in a specific frequency band, while other frequency band signals are not affected.

[0026] The τ-p domain is a discretized linear Radon transform, which cuts the seismic data in the time domain and space domain according to different intercept times τ and slopes p, and then obliquely stacks and projects them to the τ-p domain. It then performs a one-dimensional Fourier transform and maps the τ-p domain results to the fp domain. Based on the relationship between velocity and slowness, the fp domain results are mapped to the fv domain. The dispersion phenomenon of surface wave energy can be seen from the fv domain, and then the dispersion curve can be extracted based on the maximum characteristics of the energy cluster.

[0027] The τ-p method has a high resolution for high-order surface waves, and after improvement, it has the advantages of low distortion, high reliability, good suppression of false frequencies and endpoint effects for extracting transient Rayleigh wave dispersion curves.

[0028] The process of summing the amplitudes along the linear time-difference tilt path in the time-space domain is called tilt stacking, which is equivalent to decomposing the wave field into plane wave components. The summation path is: φ(τ,p,h)=τ+ph, where p is the ray parameter, p=sinθ / v, v represents the velocity, and θ represents the angle increment.

[0029] The forward-adjoint operator pair of the linear Radon transform is:

[0030] Preferably, the method for removing surface waves from seismic data further comprises the following steps: after step (4) is completed, the seismic data is reconstructed using inverse wavelet transform.

[0031] After the signal is transformed by discrete wavelet, the wavelet sequence ψ is obtained. i,k and its corresponding dual sequence Based on the above two sequences, the original signal f(t) can be inverted through the wavelet coefficients, that is, the inverse transform formula of discrete wavelet transform is: Where c is a constant, is i,k (t) is the dual sequence. When the discrete wavelet function ψ i,k When a0=2 and d0=1 in (t), the discretization expression formula of orthogonal wavelet can be obtained as follows: i,k (t) = 2 -i / 2 ψ(2 -i tk). BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Schematic diagram of a flow chart of a method for removing surface waves from seismic data according to an embodiment of the present invention;

[0033] Figure 2 Schematic diagram of a single shot record obtained after static correction and spherical diffusion compensation in an embodiment of the present invention;

[0034] Figure 3 It is a schematic diagram of the result of measuring surface wave on a single shot in an embodiment of the present invention;

[0035] Figure 4 The amplitude spectrum of a single shot is obtained in the embodiment of the present invention;

[0036] Figure 5 Schematic diagram of single shot recording in each frequency band in an embodiment of the present invention;

[0037] Figure 6 It is a schematic diagram of the seismic data with a frequency band of 0-18 Hz before and after attenuation and the attenuated noise in an embodiment of the present invention;

[0038] Figure 7 It is a schematic diagram of the seismic data with a frequency band of 18-36 Hz before and after attenuation and the attenuated noise in an embodiment of the present invention;

[0039] Figure 8It is a schematic diagram of seismic data with a frequency band of 0-36 Hz in the data processed by the conventional method in Experimental Example 1 of the present invention and seismic data with a frequency band of 0-36 Hz in the data processed by the method of the embodiment;

[0040] Fig. 9 Schematic diagram of residuals corresponding to different denoising methods in Experimental Example 1 of the present invention;

[0041] Fig.10 The amplitude spectra of the seismic data (original seismic data) before denoising and the seismic data after being processed by different denoising methods in Experimental Example 1 of the present invention;

[0042] Fig.11 A schematic diagram of a single shot record of seismic data of different frequency bands obtained by frequency division scanning of seismic data processed by a conventional denoising method in Experimental Example 1 of the present invention;

[0043] Fig.12 A schematic diagram of a single shot record of seismic data of different frequency bands obtained by frequency division scanning of seismic data processed by the method of the embodiment in Experimental Example 1 of the present invention;

[0044] Fig.13 It is a comparison diagram of the seismic data before denoising in Experimental Example 1 of the present invention, the stacked section corresponding to the seismic data after denoising by a conventional method, and the stacked section corresponding to the seismic data after being processed by the method of the embodiment;

[0045] Fig.14 The schematic diagram is a result of attenuation denoising of seismic data with a frequency band of 0-16 Hz by using the method of comparative example 1 in experimental example 2 of the present invention and a result of attenuation of seismic data with a frequency band of 0-18 Hz by using the method of the embodiment, as well as the residual results of the two;

[0046] Fig.15 The schematic diagram shows the result of attenuation denoising of seismic data with a frequency band of 16-32 Hz by using the method of comparative example 1 in experimental example 2 of the present invention and the result of attenuation of seismic data with a frequency band of 18-36 Hz by using the method of the embodiment, as well as the residual results of the two.

[0047] Fig.16 A comparison diagram of amplitude quantization data of eight sample points in the same seismic trace in the original seismic data in Experimental Example 2 of the present invention, the seismic data processed by the method of Comparative Example 1, and the seismic data processed by the method of the embodiment;

[0048] Fig.17The schematic diagram is a result of attenuation denoising of seismic data with a frequency band of 0-16 Hz by using the method of comparative example 2 in experimental example 2 of the present invention and a result of attenuation of seismic data with a frequency band of 0-18 Hz by using the method of the embodiment, as well as the residual results of the two;

[0049] Fig.18 The schematic diagram shows the result of attenuation denoising of seismic data with a frequency band of 16-32 Hz by using the method of comparative example 2 in experimental example 2 of the present invention and the result of attenuation of seismic data with a frequency band of 18-36 Hz by using the method of the embodiment, as well as the residual results of the two.

[0050] Fig.19 It is a comparison chart of amplitude quantization data of 8 sample points in the same seismic trace in the original seismic data in Experimental Example 2 of the present invention, the seismic data processed by the method of Comparative Example 2, and the seismic data processed by the method of the embodiment. DETAILED DESCRIPTION

[0051] The technical solution of the present invention is further described below in conjunction with specific embodiments.

[0052] Example

[0053] The method for removing surface waves from single shot seismic data of this embodiment is as follows: Figure 1 As shown, the specific steps include:

[0054] (1) Static correction is performed on the acquired single-shot seismic data, and then spherical diffusion compensation is applied to eliminate the influence of energy loss of seismic waves during propagation on the seismic data, so that the seismic data (amplitude) reaches a relatively stable state within the same single shot. The processed single-shot record is as follows: Figure 2 As shown;

[0055] (2) Determine the main frequency band of surface waves

[0056] According to the characteristics of noise, the frequency band range of surface waves is investigated through octave frequency scanning, the main frequency band affected by surface waves is selected, and the main frequency band of surface wave noise is separated. First, it is measured on a single gun, with the surface wave velocity of 300-900m / s as the measurement standard. The measurement result diagram is shown in the figure below. Figure 3 As shown ( Figure 3 The area surrounded by the red line is the measured surface wave), and then the amplitude spectrum of a single shot is obtained. The result is as follows Figure 4 As shown, Figure 4The horizontal axis is frequency in Hz, and the vertical axis is amplitude in dB. The two frequencies corresponding to the amplitude of -20dB in the amplitude spectrum of a single shot are the minimum value (3Hz) and the maximum value (16Hz) of the initial frequency band of the surface wave, and the initial frequency band of the surface wave is obtained to be 3-16Hz. Due to the low frequency of the surface wave and in order to meet the octave grouping principle, the minimum value of the main frequency band of the surface wave is determined to be 0Hz; in addition, in order to ensure that the surface wave is completely attenuated, the maximum value of the main frequency band of the surface wave is determined to be the sum of the maximum value of the initial frequency band of the surface wave and 2Hz. Therefore, the maximum value of the main frequency band of the surface wave is 18Hz, that is, the main frequency band of the surface wave is 0-18Hz;

[0057] (3) Using discrete wavelet transform to divide the full-band seismic data into octaves

[0058] Based on the main frequency band of the surface wave determined in step (2), the seismic data is divided into octaves by wavelet transform. The seismic data of the full frequency band is divided into 5 octaves. The frequency bands of the 5 octaves are 0-18 Hz, 18-36 Hz, 36-72 Hz, 72-144 Hz, and 144-288 Hz, respectively. The single shot records of each frequency band are as follows: Figure 5 As shown;

[0059] (4) Analysis of the seismic data of the five frequency bands divided in step (3) shows that regular interference can be seen in the single-shot records of the frequency bands of 0-18 Hz and 18-36 Hz, indicating that the seismic data of these two octaves are significantly interfered by surface waves, and the seismic data of these two frequency bands are subjected to noise suppression;

[0060] For seismic data with a frequency band of 0-18 Hz, anti-aliasing surface wave suppression technology based on elastic wave modeling is used for attenuation. The schematic diagram of seismic data before and after attenuation and the attenuated noise is shown in the figure below. Figure 6 As shown; Figure 6 In the figure, the horizontal axis is the track number and the vertical axis is the length;

[0061] For seismic data with a frequency band of 18-36 Hz, the τ-p domain high-precision refraction wave attenuation technology is used to attenuate the surface waves and refraction waves in the seismic data with a frequency band of 18-36 Hz. The schematic diagram of the seismic data before and after attenuation and the attenuated noise is shown in the figure below. Figure 7 As shown; Figure 7 In the figure, the horizontal axis is the track number and the vertical axis is the length;

[0062] (5) Reconstruct seismic data using inverse wavelet transform.

[0063] Comparative Example 1

[0064] The difference between the method for removing surface roll in single-shot seismic data of the present comparative example and the method for removing surface roll in single-shot seismic data of the embodiment is that, in the method for removing surface roll in single-shot seismic data of the present comparative example, attenuation denoising processing is performed on seismic data with frequency bands of 0-16 Hz and 16-32 Hz, and the method for attenuating the seismic data of 0-16 Hz and 16-32 Hz is respectively the same as the method for attenuating the seismic data of 0-18 Hz and 18-36 Hz in the embodiment.

[0065] Comparative Example 2

[0066] The difference between the method for removing surface roll in single-shot seismic data of the present comparative example and the method for removing surface roll in single-shot seismic data of the embodiment is that, in the method for removing surface roll in single-shot seismic data of the present comparative example, attenuation denoising processing is performed on seismic data with frequency bands of 0-20Hz and 20-40Hz, and the method for attenuating the seismic data of 0-20Hz and 20-40Hz is respectively the same as the method for attenuating the seismic data of 0-18Hz and 18-36Hz in the embodiment.

[0067] Experimental Example 1

[0068] In order to compare with the conventional denoising method, the conventional method (attenuating the full-band data by using the anti-aliasing surface wave suppression technology based on elastic wave modeling) and the method of the embodiment are respectively used to remove the surface wave of the same single-shot seismic data (the maximum value of the initial frequency band of the surface wave confirmed when the surface wave is removed by the method of the embodiment is 16 Hz). The seismic data with a frequency band of 0-36 Hz in the data processed by the conventional method and the seismic data with a frequency band of 0-36 Hz in the data processed by the method of the embodiment are listed in Figure 8 In the above, the residuals are obtained based on the seismic data after being processed by different denoising methods and the seismic data before denoising, and the residuals (attenuated noise) corresponding to different denoising methods are obtained. The results are as follows: Fig. 9 As shown in Figure 2, the amplitude spectra of the seismic data before denoising (original seismic data) and the seismic data after being processed by different denoising methods are shown in Figure 2. Fig.10 As shown ( Fig.10 The horizontal axis is frequency in Hz, and the vertical axis is amplitude).

[0069] Surface waves are low-frequency. In order to check the attenuation effect of surface waves, the frequency bands are finely divided. The attenuation effects and fidelity of different methods are evaluated by comparison. Specifically, the seismic data processed by the conventional denoising method (the anti-aliasing surface wave suppression technology based on elastic wave modeling is used to attenuate the seismic data of the full frequency band) are subjected to frequency division scanning to obtain single-shot records of seismic data in different frequency bands (frequency bands are 0-3Hz, 3-9Hz, 9-18Hz, 18-36Hz, 36-72Hz, and 72-144Hz). The results are as follows: Fig.11 At the same time, the seismic data processed by the method of the embodiment is subjected to frequency division scanning to obtain single shot records of seismic data in different frequency bands (frequency bands are 0-3Hz, 3-9Hz, 9-18Hz, 18-36Hz, 36-72Hz, and 72-144Hz). The results are shown in FIG. Fig.12 As shown in the figure, it can be seen that the method of the present invention does not remove effective signals when removing surface waves, thus achieving the effect of preserving fidelity.

[0070] Finally, the stacked sections corresponding to the seismic data before denoising, the seismic data denoised by conventional methods, and the seismic data processed by the method of the embodiment are compared and plotted. The results are as follows: Fig.13 shown.

[0071] Experimental Example 2

[0072] In order to evaluate the denoising rate and fidelity of different methods for processing noise, the method of the embodiment and the method of comparative examples 1-2 are respectively used to remove the surface wave of the same single-shot seismic data (the maximum value of the initial frequency band of the surface wave of the single-shot seismic data when the surface wave is removed by the method of the embodiment is 16 Hz). Then, the result of attenuation denoising of the seismic data with a frequency band of 0-16 Hz by the method of comparative example 1 and the result of attenuation denoising of the seismic data with a frequency band of 0-18 Hz by the method of the embodiment and the residual results of the two are listed in FIG. Fig.14 The results of attenuation denoising of seismic data with a frequency band of 16-32 Hz by the method of comparative example 1 and the results of attenuation of seismic data with a frequency band of 18-36 Hz by the method of the embodiment and the residual results of the two are listed in Fig.15 Finally, the amplitude values ​​corresponding to multiple sample points in the same seismic trace in the original seismic data, the seismic data processed by the method of Comparative Example 1, and the seismic data processed by the method of the embodiment (the time difference between any two adjacent sample points is 0.01 ms, and the numbering of each sample point is 1001, 1002, 1003, 1004, 1005, 1006, 1007, 1008, respectively) are summarized in Tables 1 and Fig.16 ( Fig.16The horizontal axis is the sample value of the seismic trace, the interval between adjacent sample values ​​is 0.01ms, and the vertical axis is the amplitude value. Fig.16 The scatter point connection line fitting diagram is obtained by connecting the amplitude values ​​corresponding to each sample point.

[0073] At the same time, the results of attenuation denoising of seismic data with a frequency band of 0-20 Hz using the method of comparative example 2 and the results of attenuation denoising of seismic data with a frequency band of 0-18 Hz using the method of the embodiment and the residual results of the two are listed in Fig.17 The results of attenuation denoising of seismic data with a frequency band of 20-40 Hz by the method of comparative example 2 and the results of attenuation of seismic data with a frequency band of 18-36 Hz by the method of the embodiment and the residual results of the two are listed in Fig.18 Finally, the amplitudes corresponding to multiple sample points in the same seismic trace in the original seismic data, the seismic data processed by the method of Comparative Example 2, and the seismic data processed by the method of the embodiment (the time difference between any two adjacent sample points is 0.01 ms, and the numbering of each sample point is 1001, 1002, 1003, 1004, 1005, 1006, 1007, 1008, respectively) are summarized in Tables 2 and 3. Fig.19 ( Fig.19 The horizontal axis is the sample value of the seismic trace, the interval between adjacent sample values ​​is 0.01ms, and the vertical axis is the amplitude value. Fig.19 The scatter point connection line fitting diagram is obtained by connecting the amplitude values ​​corresponding to each sample point.

[0074] Table 1 Amplitude values ​​corresponding to multiple sample points in the same seismic trace in the original seismic data, the seismic data processed by the method of Comparative Example 1, and the seismic data processed by the method of the embodiment

[0075]

[0076] Table 2 Amplitude values ​​corresponding to multiple sample points in the same seismic trace in the original seismic data, the seismic data processed by the method of Comparative Example 2, and the seismic data processed by the method of the embodiment

[0077]

[0078] From Table 1 and Fig.16 It can be seen that compared with the method of comparative example 1, the amplitude of the seismic data processed by the method of the embodiment is closer to the amplitude of the original seismic data, indicating that the method of the embodiment has a high denoising rate and good fidelity.

[0079] From Table 2 and Fig.19It can be seen that compared with the method of comparative example 2, the amplitude of the seismic data processed by the method of the embodiment is closer to the amplitude of the original seismic data, indicating that the method of the embodiment has a high denoising rate and good fidelity.

Claims

1. A method for removing surface waves from single-shot seismic data, characterized in that: The following steps are involved: (1) Performing static correction and spherical diffusion compensation on the single-shot seismic data to obtain processed single-shot seismic data; (2) investigating the frequency band range of the surface wave in the processed single-shot seismic data by octave frequency scanning, and determining the main frequency band of the surface wave; the main frequency band of the surface wave is 0 to x Hz, where x is equal to the sum of the maximum value of the initial frequency band of the surface wave and y, and y is 1 to 2 Hz; (3) using discrete wavelet transform to divide the processed single-shot seismic data of the full frequency band according to octaves, the divided frequency bands include a first frequency band and a second frequency band, the first frequency band is 0 to x Hz, and the second frequency band is x to 2×(x+2) Hz; (4) Anti-aliasing processing is performed on the seismic data in the first frequency band, and τ-p transformation is used to attenuate the seismic data in the second frequency band.

2. The method for removing surface waves in single shot seismic data according to claim 1, characterized in that: The minimum frequency in the initial frequency band of the surface wave is equal to the minimum value in the frequency range corresponding to the set of data points with a velocity of 300 to 900 m / s in the processed single-shot seismic data, and the maximum frequency in the initial frequency band of the surface wave is equal to the maximum value in the frequency range corresponding to the set of data points with a velocity of 300 to 900 m / s in the processed single-shot seismic data.

3. The method for removing surface waves in single shot seismic data according to claim 2, characterized in that: The initial frequency band of the surface wave is obtained by measuring the processed single-shot seismic record.

4. The method for removing surface waves in single shot seismic data according to claim 1, characterized in that: The method for removing surface waves from seismic data also includes the following steps: after step (4) is completed, the seismic data is reconstructed using inverse wavelet transform.

Images