A surface-consistent deconvolution method for wide-azimuth seismic data

Abstract

The present application relates to a kind of wide azimuth seismic data surface consistency deconvolution method.Mainly solve the problem that the existing method does not consider that the resolution will be reduced by the isotropicity and phase consistency of wavelet and then affect the resolution improvement.The method includes inputting prestack seismic data;The data is divided into common offset vector slices;According to the design analysis window range required by the target layer;Spectrum analysis is carried out on the seismic data in the designed window range;Surface consistency spectrum decomposition is carried out on the statistical spectrum analysis data, and the surface consistency spectrum decomposition data of different components is obtained;Deconvolution operator is calculated for each component of spectrum decomposition data, and deconvolution operator is applied to each input seismic trace;Output the data after surface consistency deconvolution.The method improves the resolution and fidelity of seismic data processing results, and can obtain more fine and ideal deconvolution processing effect, to provide reliable basic data for seismic data interpretation and prestack inversion.

Description

Technical Field

[0001] This invention relates to the field of seismic exploration technology, specifically to the seismic data processing stage in geophysics within this field. More specifically, this invention relates to a surface-consistent deconvolution method for wide-azimuth seismic data. Background Technology

[0002] In the detailed seismic exploration stage, favorable reservoirs are mostly thin-interbedded structures, with effective sandstone thickness typically ranging from 3 to 5 meters. The sand bodies are small in scale, and the seismic response characteristics of the reservoirs are not obvious, making effective reservoir identification challenging. Seismic data processing often employs resolution enhancement to effectively identify thin-interbedded reservoirs; however, excessive resolution enhancement can compromise data fidelity. Deconvolution technology is a crucial and effective method for improving resolution in seismic data processing. It significantly impacts the vertical resolution and signal-to-noise ratio of seismic data, enabling a more accurate and detailed description of subsurface geological structures and improving the efficiency and accuracy of seismic data interpretation.

[0003] During the propagation of seismic wavelets underground, the influence of surface conditions can be considered a filtering effect, causing changes in the frequency and phase characteristics of the seismic signal. Due to significant variations in surface conditions and inconsistencies in excitation and reception conditions, the wavelet waveforms and frequency characteristics of reflected waves differ across different locations. To eliminate these differences and ensure that the frequency variations of reflected waves better reflect the characteristics of underground lithology, laying a solid foundation for subsequent residual static correction and velocity analysis, this filtering effect must be reversed. In this case, it is generally assumed that the filtering effect at the same location on the surface is independent of the incident angle of the seismic wave; the filtering effect is the same regardless of whether the reflection is from shallow, intermediate, or deep layers. Therefore, we call this reverse filtering method surface-consistent deconvolution. After surface-consistent deconvolution processing, the changes in wavelet waveforms caused by surface inconsistencies can be eliminated, making the seismic wave waveform more uniform. Due to limitations in exploration technology and cost, previous acquisition systems were designed with narrow azimuth, resulting in limited azimuth information carried by seismic data, thus restricting the utilization of azimuth information. With the advancement of computer and acquisition technologies, more and more wide-azimuth data are available for our use. Wide-azimuth seismic acquisition provides us with richer data on the differences caused by azimuth variations. Current surface consistency deconvolution methods for pre-stack seismic data are based on the Fourier transform principle, utilizing spectral decomposition to perform spectral analysis, spectral decomposition (generally decomposed into four components: shot point, receiver point, CMP, and offset), and the application of inverse filtering factors to complete surface consistency deconvolution processing. Traditional surface consistency deconvolution methods can meet the needs of seismic data in structural exploration, but because the four components of the decomposition do not consider the influence of azimuth information on the wavelet, the wavelet is not consistent within the 360-degree common offset range. The wavelet morphology changes with azimuth, which is a significant factor affecting the excitation wavelet morphology. Ignoring azimuth information leads to the assumption that the wavelet within the common offset range is consistent, which inevitably reduces the wavelet's directional consistency and phase consistency, thus affecting resolution improvement. Summary of the Invention

[0004] This invention addresses the problem in existing methods that fail to consider azimuth information, which reduces the wavelet co-directionality and phase consistency, thus affecting resolution improvement. It provides a surface-consistent deconvolution method for wide-azimuth seismic data. This method, tailored to the characteristics of wide-azimuth data, fully considers the azimuth-related wavelet variation features caused by non-surface factors, improving the resolution and fidelity of seismic data processing results. It achieves more refined and ideal deconvolution processing effects, providing reliable foundational data for seismic data interpretation and pre-stack inversion.

[0005] The present invention solves its problem through the following technical solution: the surface consistency deconvolution method for wide-azimuth seismic data includes the following steps: S1. Input the data to be deconvolutioned; perform pre-stack preprocessing on the data to be deconvolutioned; obtain the pre-stack preprocessed data. S2. Based on the pre-stack preprocessed data, divide the shot-receiver distance vector slices; obtain the data of the divided shot-receiver distance vector slices; S3. Based on the data of the completed gun-receiver distance vector slices, design the analysis window; obtain the logarithmic spectrum or complex spectrum data within the analysis window; S4. Perform statistical analysis on the logarithmic spectrum or re-examination spectrum data within the analysis window, and perform surface consistency decomposition on the statistical data to obtain four component data; S5. Based on the four component data obtained from the decomposition, calculate the deconvolution operator for each component; multiply the obtained deconvolution operator with the data that has completed pre-stack preprocessing to obtain the surface-consistent deconvolution data, and output the surface-consistent deconvolution data.

[0006] Furthermore, step S1, which involves pre-processing the data that needs to be deconvolved, includes: Decode the raw data; Define the observation system; Set the track header for the raw data; Noise suppression is applied to the pre-stack data; Surface consistency amplitude compensation was performed on pre-stack data; The signal-to-noise ratio of the pre-stack data after noise suppression is above 1.0; After surface consistency amplitude compensation is applied to the pre-stack data, the energy of the data is uniform and consistent in both the longitudinal and transverse directions, with no amplitude distortion. The data gather types that need to be deconvolved are shot gathers, common center point gathers, and common receiver point gathers.

[0007] Furthermore, step S2, based on the pre-stack preprocessed data, uses the following method to divide the shot-receiver distance vector patch: The pre-processed data is sorted into cross-shaped gathers; Each cross-shaped gather is divided into multiple rectangles based on equal distances between the shot line distance and the detector line distance, and each rectangle is a vector slice. By extracting vector slices in the same direction from all the cross-shaped trace sets, a vector slice trace set is formed. Each vector slice trace set is a single coverage data volume that covers the entire work area.

[0008] Furthermore, for regular observation systems with uniform coverage, the number of vector slice gathers is equal to the coverage of the work area.

[0009] Furthermore, the method for designing the analysis window in step S3 is as follows: taking the marker layer of the target layer as the center, extend an appropriate time length to the shallow and deep layers respectively to form a rectangular frame, and use the rectangular frame as the analysis window for deconvolution.

[0010] Furthermore, the time duration for extending into the shallow and deep layers includes all information about the target layer.

[0011] Furthermore, the method for selecting the analysis window in step S3 includes: To avoid affecting the wavelet estimation data, we chose to avoid the initial strong energy and the strong energy near the shot point. Choose a relatively stable time period for the wavelet. While ensuring that the wavelet remains basically unchanged, the time window should be as large as possible in order to take into account the deconvolution calculations of shallow, medium and deep layers. Select a single time window; The offset and coverage number of the selected data should not be too low.

[0012] Furthermore, the maximum offset distance is greater than 2000m; the maximum number of coverages is greater than 60.

[0013] Furthermore, the data in step S4 includes the shot point, receiver point, common center point, and common offset azimuth angle. Furthermore, the common offset azimuth term is constructed from offset and azimuth information; it is a small rectangle with azimuth information obtained by dividing the pre-stack preprocessing data according to the shot distance and receiver distance.

[0014] Compared with the above-mentioned background technology, the present invention has the following beneficial effects: This invention discloses a surface consistency deconvolution method for wide-azimuth seismic data. By considering the azimuth information of the wide-azimuth data, the data is divided according to the shot line distance and receiver line distance, resulting in many small rectangles. Each small rectangle with a limited range of offset and azimuth is treated as a unit. This unit can retain the azimuth information of the wavelet and increases the factors affecting the excitation wavelet morphology. Therefore, after deconvolution processing, the wavelet consistency of the data within the common offset range is stronger, virtual reflections are compressed, and more refined and higher-fidelity broadband deconvolution data can be obtained.

[0015] Pre-stack data processed using the surface consistency deconvolution method for wide-azimuth seismic data according to this invention were compared with seismic data without deconvolution processing in terms of spectrum and wavelet consistency. The data processed using this invention showed better wavelet compression, enhanced lateral consistency, elimination of virtual reflection effects, and a wider bandwidth compared to conventional deconvolution methods. This also effectively improved data resolution and enriched inter-layer information. This invention, specifically designed for wide-azimuth data, fully considers the azimuth-related wavelet variation characteristics caused by non-surface factors, resulting in more refined and ideal deconvolution processing effects. Ultimately, it achieved good results in the application to multiple actual seismic data blocks in the Daqing Oilfield, demonstrating excellent application prospects. Attached image description: Figure 1 This is a flowchart of a surface consistency deconvolution method for wide-azimuth seismic data according to the present invention; Figure 2 This is a schematic diagram of the common-shot-detector-vector patch division method according to an embodiment of the present invention. Detailed implementation method: To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0016] This invention discloses a surface consistency deconvolution method for wide-azimuth seismic data, comprising: inputting pre-stack seismic data; dividing the data into common shot-receiver offset vector patches; designing an analysis time window range according to the target layer requirements; performing spectral analysis on the seismic data within the designed time window range; performing surface consistency spectral decomposition on the statistical spectral analysis data to obtain surface consistency spectral decomposition data of different components; calculating deconvolution operators for each component of the spectral decomposition data and applying the deconvolution operators to each input seismic trace; and outputting the surface consistency deconvolutioned data. Because this invention considers the azimuth information of wide-azimuth data, dividing the data according to shot-receiver and receiver-receiver offsets to obtain many small rectangles, each small rectangle with a limited range of offset and azimuth angle serves as a unit for convolution. This unit can retain the azimuth information of the wavelet and increases the factors affecting the excitation wavelet morphology. Therefore, after deconvolution processing, the wavelet consistency of the data within the common offset range is stronger, virtual reflections are compressed, and more refined and higher-fidelity broadband deconvolutioned data can be obtained, which has broad application and promotion potential.

[0017] like Figure 1 As shown, a surface consistency deconvolution method for wide-azimuth seismic data includes the following steps: S1. Input the data to be deconvolutioned; perform pre-processing on the data to be deconvolutioned. Pre-stack preprocessing steps include raw data unpacking, observation system definition, trace head placement, pre-stack noise suppression, and surface consistency amplitude compensation. Pre-stack noise suppression must ensure a signal-to-noise ratio above 1.0, and surface consistency amplitude compensation must ensure uniform energy across the longitudinal and transverse directions, without amplitude distortion, to guarantee the quality of the pre-stack data.

[0018] The input data must be pre-stack seismic data. There are no restrictions on the type of input data gathers, which can be shot gathers, common center point gathers, and common receiver gathers.

[0019] S2. Based on the pre-stack preprocessed data, divide the shot-receiver distance vector slices; The specific process and method for dividing the shot-receiver distance vector patch are as follows: After pre-stack processing, the data is sorted into cross-shaped gathers to obtain cross-shaped domain data. Each cross-shaped gather is divided into multiple rectangles based on equal shot distance and receiver distance, and each rectangle is a vector slice. Each vector slice has a limited range of offset and azimuth information. If the acquisition aspect ratio is 1 and the rectangles are small enough, theoretically, we can obtain vector slices in any azimuth. Vector slices in the same direction from all cross-shaped gathers are grouped together to form a vector slice gather. Vector slice gathers not only retain the offset and azimuth information of pre-stack seismic traces but also have the advantage of uniform spatial sampling. Each vector slice gather is a single-coverage data volume covering the entire observation area. For regular observation systems with uniform coverage, the number of vector slice gathers equals the number of coverages for the observation area.

[0020] S3. Based on the data from the completed gun-receiver distance vector slices, perform the design analysis window: Centered on the marker layer of the target layer, extend an appropriate time length into the shallow and deep layers respectively. The length should be such that it contains all the information of the target layer, such as 3000ms, forming a rectangular frame as the analysis window for deconvolution.

[0021] When designing the time window, it's crucial to avoid initial strong energies and near-shot point strong energies as much as possible to prevent impacting wavelet estimation data. The time window should ideally be selected within a relatively stable wavelet timeframe, and while ensuring the wavelet remains essentially unchanged, it should be as large as possible to accommodate deconvolution calculations across shallow, medium, and deep layers. Furthermore, the time window must be a single window; multi-window processing may affect the wavelet morphology. To ensure the deconvolution effect of this invention, the offset and coverage times of the data should not be too low when selecting the time window for data analysis. This invention recommends application to data with a maximum offset of 2000m or more and a maximum coverage times of more than 60 for better results.

[0022] S4. Perform statistical analysis on the logarithmic spectrum or complex spectrum data within the analysis window, and perform surface consistency decomposition on the statistical seismic data to obtain four component data. When performing surface-consistent deconvolution on the data, four components are obtained: shot point, receiver point, common center point, and common offset azimuth term. The common offset azimuth term is a small rectangle containing azimuth information, obtained by dividing the actual data according to the shot line distance and receiver line distance. Since the wavelet shape of actual seismic data is affected not only by surface consistency factors but also by subsurface non-surface consistency factors, the more terms decomposed, the better the wavelet consistency effect. It is important to note that the number of terms decomposed should correspond to the number of terms selected for deconvolution to achieve satisfactory results.

[0023] S5. Based on the four component data obtained from statistical decomposition, calculate the deconvolution operator for each component; multiply the obtained deconvolution operator with the data after pre-stack preprocessing to obtain the surface-consistent deconvolution data, and output the surface-consistent deconvolution data.

[0024] This step supports pre-stack seismic data, and the gather type is not limited; it can be shot gathers or common midpoint gathers.

[0025] This invention takes into full account the azimuth-related wavelet variation characteristics caused by non-surface factors, based on the characteristics of wide azimuth data, and can obtain a more refined and ideal deconvolution processing effect.

[0026] Example 1 This invention discloses a surface-consistent deconvolution method for wide-azimuth seismic data, illustrated by its application in actual data testing in the L exploration area of ​​the northern Songliao Basin. This area is rich in oil and gas resources and is one of the important exploration areas in the northern Songliao Basin. The exploration target is the thin, narrow, and small-channel sand bodies of the Fuyu oil layer. Depicting these small target bodies places higher demands on the resolution of the seismic data. The data for this area was acquired in a wide azimuth with an aspect ratio of 0.8 and shot and receiver spacing of 160m. After preprocessing including data decoding, observation system definition, track head placement, pre-stack noise suppression, and surface-consistent amplitude compensation, analysis shows that the overall signal-to-noise ratio of the data is between 1.3 and 3.7. Analysis of the overall energy plane map shows that the energy variation in the data is uniform and there are no abrupt changes. Overall, the data quality meets the requirements of the deconvolution processing method of this invention.

[0027] S1. Perform quality control on the data after completing pre-stack processing such as noise suppression and surface uniformity amplitude compensation. Check whether the signal-to-noise ratio and energy uniformity of the data meet the requirements and prepare for deconvolution processing. S2. Divide the shot-receiver distance vector slices based on the pre-stack preprocessed data; The specific process and method for dividing the shot-receiver distance vector patch is as follows: First, the preprocessed data is sorted into cross-shaped gathers to obtain the data of the cross-shaped domain. Each cross-shaped gather is then divided into multiple small rectangles at equal intervals based on the shot-to-receiver distance of 160m and the receiver-to-receiver distance of 160m, such as... Figure 2 As shown in the diagram, the vertical axis represents the shot line and the horizontal axis represents the receiver line. The data was divided according to the shot line distance and receiver line distance, forming multiple small rectangles with a side length of 160m. Each small rectangle can be considered as a shot-receiver distance vector sheet unit with a finite range of offset and azimuth angle, and this unit can retain azimuth information. Vector sheets with the same azimuth from all the cross-shaped gathers are grouped together to form a vector sheet gather. Sorting all vector sheets with the same azimuth completes the vector sheet division of the data. In the actual data, there are 187 vector sheet gathers, approximately equal to 189 coverage times for the work area.

[0028] S3. Design analysis window for the data after the gun-receiver distance vector patch division: Centered on the target layer T2, extend 1500ms into both the shallow and deep layers to include all data within the target layer, forming a rectangular frame from 500ms to 3500ms as the analysis window for deconvolution.

[0029] S4. Perform spectral statistics and unified surface consistency decomposition on the data within the analysis window to obtain four component data: shot point, receiver point, common center point, and common offset azimuth.

[0030] S5. Based on the four component data obtained from the statistical decomposition of the composite spectrum, deconvolution operators are calculated for each component. The obtained deconvolution operators are then multiplied with the pre-stack data to obtain the surface-consistent deconvolutioned data. This processing was performed at the shot gather. Quality control checks on the shot gather data after deconvolution processing revealed that the hyperbolic characteristics of the reflection axes were more prominent after deconvolution, the wavelet was better compressed, and the lateral consistency was enhanced. Under the same conditions, by comparing and analyzing the data obtained by the conventional deconvolution method, spectral analysis of the target layer showed that the data processed by the method of this invention had a bandwidth broadened by 3-5 Hz compared to the conventional method. The autocorrelation analysis of the wavelet revealed that the autocorrelation value was improved from the conventional distribution of 6-8 to a distribution of 5-6, and the lateral consistency was also better than the conventional method. This proves that the present invention has advantages in compressing the wavelet and improving resolution, and also lays a good data foundation for subsequent azimuth processing and pre-stack inversion.

Claims

1. A surface-consistent deconvolution method for wide-azimuth seismic data, characterized in that: Includes the following steps: S1. Input the data that needs to be deconvolved; Perform pre-stack preprocessing on the data that needs to be deconvolved; Obtain the pre-processed data before stacking; S2. Based on the pre-stack preprocessed data, divide the shot-receiver distance vector slices; obtain the data of the divided shot-receiver distance vector slices; S3. Based on the data of the completed gun-receiver distance vector slices, design the analysis window; obtain the logarithmic spectrum or complex spectrum data within the analysis window; S4. Perform statistical analysis on the logarithmic spectrum or re-examination spectrum data within the analysis window, and perform surface consistency decomposition on the statistical data to obtain four component data; S5. Based on the four component data obtained from the decomposition, calculate the deconvolution operator for each component; multiply the obtained deconvolution operator with the data that has completed pre-stack preprocessing to obtain the surface-consistent deconvolution data, and output the surface-consistent deconvolution data. The data of the four components in step S4 include the shot point, receiver point, common center point, and common offset azimuth angle. The common offset azimuth is constructed from offset and azimuth information; it is a small rectangle with azimuth information obtained by dividing the pre-stack preprocessing data according to the shot distance and detector distance.

2. The surface consistency deconvolution method for wide-azimuth seismic data according to claim 1, characterized in that: Step S1, which involves pre-processing the data before deconvolution, includes: Decode the raw data; Define the observation system; Set the track header for the raw data; Noise suppression is applied to the pre-stack data; Surface consistency amplitude compensation was performed on pre-stack data; And / or, The signal-to-noise ratio of the pre-stack data after noise suppression is above 1.0; After surface consistency amplitude compensation is applied to the pre-stack data, the energy of the data is uniform in both the longitudinal and transverse directions. The data gather types that need to be deconvolved are shot gathers, common center point gathers, and common receiver point gathers.

3. The surface consistency deconvolution method for wide-azimuth seismic data according to claim 1, characterized in that: Step S2, based on the pre-stack preprocessed data, uses the following method to divide the shot-receiver distance vector patch: The pre-processed data is sorted into cross-shaped gathers; Each cross-shaped gather is divided into multiple rectangles based on equal distances between the shot line distance and the detector line distance, and each rectangle is a vector slice. By extracting vector slices in the same direction from all the cross-shaped trace sets, a vector slice trace set is formed. Each vector slice trace set is a single coverage data volume that covers the entire work area.

4. The surface consistency deconvolution method for wide-azimuth seismic data according to claim 3, characterized in that: For a regular observation system with uniform coverage, the number of vector slice gathers is equal to the number of coverages of the work area.

5. The surface consistency deconvolution method for wide-azimuth seismic data according to claim 1, characterized in that: The method for designing the analysis window in step S3 is as follows: taking the marker layer of the target layer as the center, extend an appropriate time length to the shallow and deep layers respectively to form a rectangular frame, and use the rectangular frame as the analysis window for deconvolution.

6. The surface consistency deconvolution method for wide-azimuth seismic data according to claim 5, characterized in that: The time taken to extend into both shallow and deep layers contains all the information of the target layer.

7. The surface consistency deconvolution method for wide-azimuth seismic data according to claim 1, characterized in that: Step S3 involves the method for selecting the analysis window, including: To avoid affecting the wavelet estimation data, we chose to avoid the initial strong energy and the strong energy near the shot point. Choose a relatively stable time period for the wavelet. While ensuring that the wavelet remains basically unchanged, the time window should be as large as possible in order to take into account the deconvolution calculations of shallow, medium and deep layers. Select a single time window; The offset and coverage number of the selected data should not be too low.

8. The surface consistency deconvolution method for wide-azimuth seismic data according to claim 7, characterized in that: The maximum offset distance is over 2000m; the maximum number of coverages is over 60.

Images