Shallow-water seabed wide-angle multiple suppression method for ocean multichannel seismic data

Abstract

The invention discloses an ocean multichannel seismic data shallow-water seabed wide-angle multiple suppression method, which comprises the following steps of: according to the characteristic that seabed wide-angle multiples are linear in middle and far offset distances, separating a primary wave from a wide-angle multiple wave field by applying linear tau-p conversion to obtain a wide-angle multiple wave model, and then subtracting from the data to obtain a subtraction result; and the purpose of suppressing wide-angle multiple waves is achieved. In the suppressing process, the suppressing range of a far and middle offset triangular area needs to be calculated through the seabed depth, and multiple waves, predictable by a conventional method, of a small offset area are not removed. According to a traditional method, seabed wide-angle multiples are cut off, medium-far offset effective signals can be damaged, the problem is solved through the scheme, only the seabed wide-angle multiples are suppressed, and the effective signals are not damaged. Meanwhile, the algorithm is simple in implementation process, high in precision, suitable for most of ocean multichannel seismic data and high in practical application and popularization value.

Description

Technical Field

[0001] The invention belongs to the field of seismic data processing, and in particular relates to a method for suppressing wide-angle multiple waves of shallow-water seabed multi-channel seismic data in ocean. Background Art

[0002] Seismic exploration is the exploration of energy such as oil and gas resources, mineral energy, and hydrate resources by using the physical property differences of underground media such as velocity and density. It can also be used to study the distribution of the earth's crust and regional geological conditions. Seismic exploration can be divided into land exploration and marine exploration. The difficulties in seismic data processing are different. The difficulty in land seismic data processing is the establishment of near-surface velocity models and static correction; the difficulty in marine seismic data processing is the suppression of multiple waves. In marine seismic data, shallow water seismic data has more types of multiple waves than deep water seismic data, and the difficulty of suppression is also higher than that of deep water seismic data.

[0003] There is a strong energy triangle near the direct wave of shallow marine multi-channel seismic data, which is characterized by a short duration near the offset and a longer duration as the offset increases; it also has the characteristics of high frequency, strong energy, and large amplitude, which masks the primary reflection wave signal. In the past, it was generally believed that the linear interference wave in the strong energy triangle was the direct wave and its dispersion. By studying the data from the continental shelf to the slope break zone, it was found that ( Fig. 9 ): The seismic wave field of the black triangle area with strong energy in shallow water is a wide-angle multiple wave of the seabed with medium and long offset and large angle ( Fig.10 ). Observe the single shot at the deep water slope ( Fig.11 ), it can be found that the time difference between the seabed reflection and its first-order multiple waves is large at near offset and becomes smaller at far offset. As the water depth decreases, the time difference between the first wave and the multiple waves at medium and far offsets becomes smaller and intertwined, forming a black triangle area with strong energy ( Fig.10 ).

[0004] Wide-angle multiple waves on the shallow seabed seriously affect the extraction and migration imaging of mid-to-far offset signals in long-array seismic data. Conventional methods often cut off seismic data outside the wide-angle multiple waves on the seabed, which goes against the original intention of using long-array seismic acquisition and causes data waste. More importantly, mid-to-far offset seismic data are of great significance for seismic data imaging with strong shielding layers in shallow areas.

[0005] For example, in a typical shallow water environment area, there is a strong reflection interface formed by the Cenozoic terrestrial clastic rock formation directly covering the Middle Paleozoic marine carbonate rock formation in the shallow part of the stratum. This interface reflects most of the energy arriving back to the sea surface, and oscillates back and forth between the sea surface, the seabed and the interface to form multi-order strong energy multiple waves. At the same time, the effective signal reaching the target layer below this layer is weak. The strong energy of the multiple waves and the weak energy of the effective waves result in poor seismic imaging quality below the shallow water strong shielding layer, thus restricting the progress of oil and gas exploration. The medium-long offset seismic data is less affected by the ocean multiple waves, has a high signal-to-noise ratio, and is sensitive to velocity changes. The medium-long offset seismic data can improve the accuracy of velocity analysis and the number of times the seismic data is covered. It is of great significance for the processing of seismic data in shallow water environments with strong shielding layers.

[0006] Therefore, a seismic data processing method that suppresses wide-angle multiple waves in shallow water seabed without damaging effective reflection signals is studied to improve the imaging quality of long-array ocean streamer seismic data in shallow water environment with strong shielding layer conditions. Summary of the invention

[0007] The present invention solves the problem that strong energy seabed wide-angle multiple waves of shallow water long-array seismic data affect imaging of medium-long offset data. A method for suppressing shallow water seabed wide-angle multiple waves of medium-long offset marine multi-channel seismic data is proposed. The method separates reflected waves from seabed wide-angle multiple waves to obtain a seabed wide-angle multiple wave model, and then subtracts the wide-angle multiple wave model from the original data, thereby achieving the purpose of suppressing shallow water seabed wide-angle multiple waves. The implementation process is simple and applicable to most shallow water multi-channel seismic data.

[0008] The present invention is implemented by adopting the following technical scheme: a method for suppressing wide-angle multiple waves of shallow-water seabed in marine multi-channel seismic data, comprising the following steps:

[0009] Step A, perform dynamic time difference correction on the original data, then pick up the seafloor reflection time on the zero offset profile, and calculate the suppression range of the seafloor wide-angle multiple waves based on it;

[0010] Step B, performing a linear τ-p transformation on the data within the suppression range, identifying the seabed wide-angle multiple waves and the primary reflection waves according to the characteristics that the reflection waves appear as ellipses and the linear waves appear as points in the τ-p domain, removing the primary reflection waves by a removal or filtering method, and separating the seabed wide-angle multiple waves and the reflection waves, and after removing the reflection waves, performing an inverse transformation back to the time-space domain to obtain a wide-angle multiple wave model;

[0011] Step C, subtracting the wide-angle multiple wave model from the original data to obtain the result data after the wide-angle multiple wave suppression of the shallow seabed.

[0012] Furthermore, in step A, the original data is required to have been loaded into the observation system. First, the minimum offset data of each shot of the original data is extracted to form the first profile; then, the seawater velocity is used to perform dynamic calibration time difference correction; and then the seabed reflection time is picked up.

[0013] Furthermore, in step A, the seafloor wide-angle multiple wave suppression range is calculated according to the seafloor reflection time and water velocity:

[0014]

[0015] Where x is the offset; t x is the time at the x offset; t x0 is the picked zero-offset seafloor reflection time; v min is the minimum speed in the suppression zone; v max is the maximum speed of the pressing area; according to the obtained t x , remove the original data and keep only t x Data within the range.

[0016] Furthermore, in step B, the following method is used:

[0017] Perform a linear τ-p transformation on the data within the suppression range, and transform the (t, x) in the spatiotemporal domain to the (τ, p) in the τ-p domain using the following formula:

[0018]

[0019] Then, the area with non-point features is cut off in the τ-p domain, that is, the reflected waves within the suppression range are cut off; after the reflected waves are cut off, the τ-p inverse transformation is performed to transform (τ, p) in the τ-p domain back to (t, x) in the space-time domain. The formula is as follows:

[0020]

[0021] Furthermore, in step C, the wide-angle multiple wave model is subtracted from the original data using the least squares method, and matched filtering is performed in time or in time and space based on the best match between the wide-angle multiple wave model and the original data under the least squares criterion, and the obtained desired filter operator is applied to the wide-angle multiple wave model, and then subtracted from the original data. The matched filtering specifically includes two methods:

[0022] (1) Matching is performed only in time: each data channel in the wide-angle multiple wave model data is autocorrelated, and all autocorrelation results are added in the current time window; then each data channel of the wide-angle multiple wave model data is cross-correlated with the corresponding original data, and all cross-correlation results are added in the current time window; finally, the Levinson recursive algorithm is used to solve the normal equation AF=C, where A is the Toeplitz matrix of the autocorrelation sum; F is the expected filter operator; and C is the sum of the cross-correlations.

[0023] (2) Matching is performed in both time and space: in the current time window, the current channel in the wide-angle multiple wave model data is autocorrelated with other channels at different time and space delays; in the current time window, each channel in the wide-angle multiple wave model data is cross-correlated with the original data at different time and space delays; the conjugate gradient method is used to solve the equation A'F = C', where A' is the calculated autocorrelation matrix; F is the expected filter operator; and C' is the calculated cross-correlation.

[0024] Compared with the prior art, the advantages and positive effects of the present invention are:

[0025] This scheme first uses the seawater velocity to perform dynamic correction on the original data, then picks up the zero-offset reflection time of the seabed, and calculates the suppression range of the seabed wide-angle multiple waves based on this, and does not predict the multiple waves of the near-medium offset data that can be predicted by the conventional multiple wave prediction method, to ensure that the seismic signals of the near-medium offset are not damaged; then the data within the suppression range is transformed into the τ-p domain, and the linear seabed wide-angle multiple waves are transformed into points in the τ-p domain, and the reflection waves with hyperbolic characteristics are ellipses in the τ-p domain. The seabed wide-angle multiple waves and the reflection waves are separated, and after cutting off the reflection waves, the inverse transformation is returned to the time-space domain to obtain the seabed wide-angle multiple wave model; finally, the multiple waves are subtracted from the original data, and the suppression of the seabed wide-angle multiple waves is achieved;

[0026] The traditional method cuts off the seabed wide-angle multiple waves and other data, resulting in data waste and reducing the number of data coverage. At the same time, the medium-long offset information is more sensitive to velocity, which is more conducive to improving the accuracy of velocity analysis. The present method predicts the seabed wide-angle multiple waves of medium-long offset and then subtracts them from the original data without damaging the medium-long offset reflection wave signals. After suppressing the seabed wide-angle multiple waves, the medium-long offset reflection waves with high signal-to-noise ratio can be used for high-precision velocity analysis and offset stacking, thereby improving the imaging effect of long-array seismic data. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A schematic diagram of a process of the method according to an embodiment of the present invention;

[0028] Figure 2 The original single shot data of the embodiment of the present invention;

[0029] Figure 3 This is the first profile after water velocity correction in the embodiment of the present invention, i.e., the picking diagram of the seabed reflection time;

[0030] Figure 4 The seafloor wide-angle multiple wave suppression range calculated by the embodiment of the present invention;

[0031] Figure 5 is the data after τ-p positive transformation in the embodiment of the present invention;

[0032] Figure 6 The data after the τ-p domain reflected wave is removed in the embodiment of the present invention;

[0033] Figure 7 The multiple wave model data obtained after τ-p inverse transformation in the embodiment of the present invention;

[0034] Figure 8 The single shot data after the seabed wide-angle multiple wave suppression obtained by the example of the present invention;

[0035] Fig. 9 It is the shot gathering data from the continental shelf to the slope break zone;

[0036] Fig.10 This is the shallow seabed wide-angle multiple wave single shot data;

[0037] Fig.11 This is the single shot data for deeper water. DETAILED DESCRIPTION

[0038] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described below in conjunction with the accompanying drawings and embodiments. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0039] For shallow-water marine multi-channel seismic data, the interference of wide-angle multiple waves on the seabed is unavoidable, and its existence affects the imaging and inversion of medium-long offset data. In order to effectively suppress the wide-angle multiple waves on the seabed without damaging the effective reflection information, this embodiment proposes a method for suppressing wide-angle multiple waves on the shallow-water seabed of marine multi-channel seismic data. The wide-angle multiple waves on the shallow-water seabed are linear at medium-long offsets, and the reflected waves are hyperbolic. The signals of the two are intertwined in the time and space domain and cannot be separated. After transforming the time and space domain data into the τ-p domain by Fourier transform, the wide-angle multiple waves on the shallow-water seabed become discrete points, and the reflected waves become ellipses, so the two can be separated to obtain a wide-angle multiple wave model. Finally, the least squares method is used to subtract the multiple waves from the original data to achieve the purpose of suppressing the wide-angle multiple waves on the shallow-water seabed, such as Figure 1 As shown, the following steps are included:

[0040] Step A, perform dynamic time difference correction on the original data, then pick up the seafloor reflection time on the zero offset profile, and calculate the suppression range of the seafloor wide-angle multiple waves based on it;

[0041] Step B, transforming the data within the suppression range into the τ-p domain, using the characteristics that the linear seafloor wide-angle multiple waves become points in the τ-p domain and the hyperbolic reflection waves are ellipses in the τ-p domain, separating the seafloor wide-angle multiple waves from the reflection waves, removing the reflection waves, and then inversely transforming back to the time-space domain to obtain the seafloor wide-angle multiple wave model;

[0042] Step C: Finally, the wide-angle multiple wave model is subtracted from the original data to obtain the result data after suppressing the wide-angle multiple wave model.

[0043] Specifically, the following is combined Figure 2 and Figure 7 The method of the present invention is described in detail:

[0044] In step A, when picking up the seabed reflection time, the original data is required to have been loaded with the observation system, and at least contain the following header information: shot point coordinates, detection point coordinates, common center point coordinates, shot number, channel number, offset; then extract the data of the minimum offset distance (i.e., the first channel) of each shot of the original data to form the first channel profile. For marine seismic data, the offset distance of each seismic data of the first channel profile is the minimum offset distance of the data; use the seawater velocity to perform dynamic calibration time difference correction on the first channel profile to eliminate the influence of the offset distance on the reflection time. At this time, the reflection time of the seabed reflects the zero offset distance reflection time. Pick up the seabed reflection time on the obtained profile and record it. The seabed reflection time can pick up the peak, trough or zero phase time, but it should be noted that the selected standard should not be changed at will.

[0045] Then the suppression range of the seabed wide-angle multiple waves is calculated based on the seabed reflection time and water speed:

[0046]

[0047] Where x is the offset; t x is the time at the x offset; t x0 is the picked zero-offset seafloor reflection time; v min is the minimum speed in the suppression zone; v max is the maximum speed of the pressing area; according to the obtained t x , remove the original data and keep only t x The data in the range is obtained according to the t x , remove the original data and keep only t Data within the range of x. Among them, v min and v maxThe selection of is related to the water depth and maximum offset of the actual data. When the water depth is 70m and the maximum offset is 7200m, v min Take 1200m / s, v max Take 1900m / s. When the water depth is greater than 70m, v min Can be increased appropriately; offset distance is less than 7200m, v max Can be reduced appropriately.

[0048] In step B, a linear τ-p transformation is performed on the data within the suppression range, and the formula for transforming (t, x) in the spatiotemporal domain to (τ, p) in the τ-p domain is as follows:

[0049]

[0050] The theoretical record of a point in the space-time domain is a straight line after the τ-p transformation; a linear wave in the space-time domain is a point after the τ-p transformation; a reflected wave in the space-time domain is an ellipse after the τ-p transformation. This feature is used to separate and remove the reflected wave group from the linear wave group in the τ-p domain, and remove the area with non-point features, that is, remove the reflected wave within the suppression range. After removing the reflected wave, perform a τ-p inverse transformation back to the space-time domain to obtain a wide-angle multiple wave model: The formula for inversely transforming (τ, p) in the τ-p domain back to (t, x) in the space-time domain is as follows:

[0051]

[0052] The seabed wide-angle multiple wave model is obtained through the above inverse transformation.

[0053] Step C, finally subtracting the wide-angle multiple wave model from the original data, that is, obtaining the data after the wide-angle multiple wave suppression of the seabed; specifically, the wide-angle multiple wave model is subtracted from the original data by the least squares method, and based on the best matching of the wide-angle multiple wave model and the original data under the least squares criterion, matched filtering is performed in time or in time and space, and the obtained expected filter operator is applied to the wide-angle multiple wave model, and then it is subtracted from the original data. The matched filtering specifically includes two methods:

[0054] (1) Matching is performed only in time: each data channel in the wide-angle multiple wave model data is autocorrelated, and all autocorrelation results are added in the current time window; then each data channel of the wide-angle multiple wave model data is cross-correlated with the corresponding original data, and all cross-correlation results are added in the current time window; finally, the Levinson recursive algorithm is used to solve the normal equation AF=C, where A is the Toeplitz matrix of the autocorrelation sum; F is the expected filter operator; and C is the sum of the cross-correlations.

[0055] (2) Matching is performed in both time and space: in the current time window, the current channel in the wide-angle multiple wave model data is autocorrelated with other channels at different time and space delays; in the current time window, each channel in the wide-angle multiple wave model data is cross-correlated with the original data at different time and space delays; the conjugate gradient method is used to solve the equation A'F = C', where A' is the calculated autocorrelation matrix; F is the expected filter operator; and C' is the calculated cross-correlation.

[0056] In order to verify the effectiveness and practicality of the present invention, some actual seismic data is selected for filtering application, and the specific processing process is as follows:

[0057] Based on the actual seismic data, the wide-angle multiple waves of the seafloor are suppressed. Figure 2 is the original shot gather data, Figure 3 The first profile extracted is corrected for water velocity and the seabed reflection time is picked up, in this case the wave crest is picked up. Figure 4 is the seafloor wide-angle multiple wave suppression range calculated according to Formula 1.

[0058] Will Figure 4 The spatial and temporal data is transformed into the τ-p domain as Figure 5 As shown, Figure 5 Reflection wave removal Figure 6 As shown, Figure 6 Perform the inverse τ-p transformation back to the space-time domain as Figure 7 As shown. Figure 4 and Figure 7 It can be seen that the reflected wave with hyperbolic characteristics is protected. Figure 7 Only wide-angle multiple waves with linear characteristics are retained.

[0059] Finally, use Figure 2 Subtract the original data Figure 7 The multiple wave model of the seabed wide-angle multiple wave suppression is obtained as follows Figure 8 It can be seen that the method of the present invention only suppresses the linear wide-angle multiple waves of the seabed, and the reflected waves masked by them are not destroyed. At the same time, the reflected waves with relatively weak energy are highlighted, and the information of medium-long offset distances can be effectively utilized.

[0060] The above description is only a preferred embodiment of the present invention and does not limit the present invention in other forms. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for suppressing wide-angle multiple waves in shallow water seabed of marine multi-channel seismic data, characterized in that: The following steps are involved: Step A, performing a dynamic time difference correction on the original data, and then picking up the seafloor reflection time on the zero offset profile, and calculating the suppression range of the seafloor wide-angle multiple waves based on it; Step B, performing a linear τ-p transformation on the data within the suppression range, cutting off the reflected wave according to the feature that the reflected wave appears as an ellipse and the linear wave appears as a point in the τ-p domain, separating the seabed wide-angle multiple waves from the reflected wave, and after cutting off the reflected wave, inversely transforming back to the time-space domain to obtain a wide-angle multiple wave model; Step C, subtracting the wide-angle multiple wave model from the original data to obtain the result data after the wide-angle multiple wave suppression of the shallow seabed.

2. The method for suppressing wide-angle multiple waves in shallow water seabed of marine multi-channel seismic data according to claim 1, characterized in that: In step A, when picking up the seabed reflection time, it is required that the original data has been loaded into the observation system, and the included header information includes: shot point coordinates, detection point coordinates, common center point coordinates, shot number, track number, offset, and then the minimum offset data of each shot in the original data is extracted to form the first track profile, and the seawater velocity is used to perform dynamic calibration time difference correction. At this time, the reflection time of the seabed reflects the zero offset reflection time.

3. The method for suppressing wide-angle multiple waves in shallow water seabed of marine multi-channel seismic data according to claim 1, characterized in that: In step A, the seafloor wide-angle multiple wave suppression range is calculated based on the seafloor reflection time and water velocity: Where x is the offset; t x is the time at the x offset; t x0 is the picked zero-offset seafloor reflection time; v min is the minimum speed in the suppression zone; v max is the maximum speed of the pressing area; according to the obtained t x The original data is cut off and only t is retained x Data within the range.

4. The method for suppressing wide-angle multiple waves in shallow water seabed of marine multi-channel seismic data according to claim 1, characterized in that: In step B, the data within the suppression range is subjected to a linear τ-p transformation, and the (t, x) in the time-space domain is transformed into the (τ, p) in the τ-p domain, and the region with non-point features is cut off in the τ-p domain, that is, the reflected waves within the suppression range are cut off; after the reflected waves are cut off, a τ-p inverse transformation is performed back to the time-space domain to obtain a wide-angle multiple wave model; wherein: τ-p transformation: Inverse τ-p transform:

5. The method for suppressing wide-angle multiple waves in shallow water seabed of marine multi-channel seismic data according to claim 1, characterized in that: In step C, the wide-angle multiple wave model is subtracted from the original data using the least square method, matched filtering is performed in time or in time and space, the obtained desired filter operator is applied to the wide-angle multiple wave model, and then it is subtracted from the original data. The matched filtering specifically includes two methods: (1) Matching only in time: each data channel in the wide-angle multiple wave model data is autocorrelated, and all autocorrelation results are added in the current time window; Then, each channel of the wide-angle multiple wave model data is cross-correlated with the corresponding original data, and all cross-correlation results are added in the current time window; finally, the Levinson recursive algorithm is used to solve the normal equation AF=C, where A is the Toeplitz matrix of the autocorrelation sum; F is the desired filter operator; and C is the sum of the cross-correlations; (2) Matching is performed in both time and space: in the current time window, the current trace in the wide-angle multiple wave model data is autocorrelated with other traces at different time and space delays; in the current time window, each trace in the wide-angle multiple wave model data is cross-correlated with the original data at different time and space delays; the conjugate gradient method is used to solve the equation A'F = C', where A' is the calculated autocorrelation matrix; F is the expected filter operator; C' is the calculated cross-correlation.

Images