Method of correcting tensile distortion by compression

Abstract

The present invention discloses a method for compressive strain correction. After pre-stack data preprocessing and common-center point domain data rearrangement, acquired seismic data is combined with well-seismic data to extract a high-fidelity wavelet from near-offset data as a target wavelet. The target wavelet is then deconvolved with a wavelet to be shaped extracted from far-offset gather data after NMO strain correction to obtain a filter operator f(t). This filter operator is then applied to the pre-stack gathers via convolution to compressive strain correction. The present invention is used in petroleum seismic exploration data analysis and processing technology to improve the imaging quality and reliability of seismic data by suppressing NMO strain distortion in far-offset data, thereby increasing the utilization rate of mid- and shallow-seismic data.

Description

Technical Field

[0001] The invention belongs to the technical field of petroleum seismic exploration data analysis and processing, and relates to dynamic correction tensile distortion processing, in particular to a method for compressive braking correction tensile distortion. Background Art

[0002] With the advancement of seismic exploration technology, the acquisition of long-offset seismic data has become the new norm. This data can be subject to severe NMO stretching during subsequent processing. Eliminating NMO stretching is crucial for improving seismic data resolution, fidelity, and shallow-sediment imaging quality. Currently, methods such as NMO gather removal, high-order NMO correction, and NMO processing are commonly used to mitigate the effects of NMO stretching. While these methods can mitigate NMO stretching to a certain extent, they also have their own drawbacks.

[0003] Eliminating NMO stretch distortion through NMO gather excision can easily lead to the loss of effective shallow reflection information, which is detrimental to shallow-to-mid-layer imaging. Furthermore, NMO stretch distortion elimination methods based on high-order NMO employ a higher-order binomial expansion. While achieving more accurate NMO velocities, while partially overcoming imaging distortion caused by NMO stretch, due to the inherent characteristics of NMO stretch, this method can only mitigate the higher-order errors in NMO moveout introduced by long-offset seismic data acquisition and cannot fundamentally suppress stretch distortion.

[0004] Conventional dynamic correction (NMO) processing can cause waveform distortion in large-offset data, reducing the frequency band, thereby reducing the resolution of thin interbeds and the accuracy of subsequent attribute analysis, adversely impacting high-precision seismic exploration. During NMO processing, wavelet stretching occurs, reducing the wavelet frequency. Correcting this stretching distortion through wavelet shaping has become an unresolved issue. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for compressive braking and stretching distortion correction, which suppresses the stretching distortion of far-offset data through wavelet extraction, deconvolution processing and the application of shaping operators, thereby achieving the imaging distortion caused by compressive braking and stretching correction, thereby achieving the purpose of improving the imaging quality and reliability of seismic data and increasing the utilization rate of medium and shallow data.

[0006] To achieve the above object, the technical solutions adopted by the present invention are as follows:

[0007] A method for compressive correction of tensile distortion, which involves pre-stack data preprocessing and common center point domain data rearrangement of seismic data, and then using well-seismic data to process near- and far-offset data to extract target wavelet s1(t) and wavelet to be shaped s2(t);

[0008] Perform deconvolution on the target wavelet s1(t) and the wavelet to be shaped s2(t) to obtain the filtering operator f(t);

[0009] The filter operator f(t) is applied to the pre-stack gathers through convolution processing to correct the tensile distortion.

[0010] As a limitation, the pre-stack pre-processed data is: seismic data is acquired and pre-processed by pre-stack denoising, amplitude compensation and deconvolution;

[0011] The common center point domain data rearrangement is: performing common center point domain data rearrangement on the pre-processed data to obtain common center point gather data.

[0012] As a further limitation, in the near-offset and far-offset data processing, the method for processing the near-offset data is:

[0013] From the shot check point offset data, the near-offset data in the common center point gather data is selected. After dynamic correction and stacking processing, the near-track stacking data is obtained. The waveform consistency is then compared with the drilling and logging synthetic record or VSP corridor stacking result:

[0014] If the waveform fit comparison result is high, the near-offset stacking data is output, and the near-track autocorrelation wavelet is extracted from the near-offset stacking data as the target wavelet s1(t);

[0015] If the waveform fit comparison result is low, re-screen the data.

[0016] The degree of agreement is considered high if the waveform, frequency, and phase consistency in the waveform comparison results are high, or if the waveform similarity coefficient between the seismic data and the drilling and logging data reaches 0.85 or above. This can be adjusted based on actual conditions according to common sense of those skilled in the art.

[0017] As a further limitation, in the near-offset and far-offset data processing, the far-offset data processing method is:

[0018] From the offset data of the shot check points, the far-offset data with stretched distortion corrected by dynamic movement correction in the common center point gather data are screened out. After dynamic movement correction and stacking processing, the far-offset stacked data are output. The far-path autocorrelation wavelet is extracted from the far-offset stacked data as the wavelet to be shaped s2(t).

[0019] The wavelet to be shaped s2(t) is a far-offset stacked data wavelet that has been corrected for stretching distortion.

[0020] Among them, the far-offset data that has been corrected for stretch distortion refers to where Δf is the difference between the dominant frequency of the common-center gather at far-offset and near-offset for the same reflection event, and f is the dominant frequency of the common-center gather at near-offset. This can be adjusted based on actual conditions according to common sense of those skilled in the art.

[0021] The deconvolution process is to solve the filter operator f(t) according to the least squares criterion and the overdetermined equation.

[0022] The deconvolution process is as follows:

[0023] (1) Determine the linear equations for the filter operator f(t)

[0024] Assuming that the residual between the output f(t)*s2(t) and the target wavelet s1(t) after the filter operator f(t) acts is e(t), then we have

[0025] e(t)=f(t)*s2(t)-s1(t).

[0026] In actual calculations, the filter operator gradually approaches 0 as time t increases infinitely. The filter operator has finite energy and is a realizable filter operator.

[0027] According to the least squares criterion, let the error function Solve the filter operator f(t), using the necessary conditions of the limit We get a linear equation system about f(t), where N and m are the lengths of the filter operator:

[0028]

[0029] (2) Obtain the filter operator f(t) by solving the overdetermined equation

[0030] First, solve the overdetermined equation R T RA=R T X, where

[0031]

[0032] Target wavelet matrix

[0033] Filter operator matrix

[0034] available,

[0035]

[0036] According to the definition of autocorrelation: x(t) is a measure of the similarity at different times in a time series.

[0037] Expressed as follows

[0038]

[0039] but,

[0040] By analogy, we can conclude that:

[0041]

[0042] At the same time, the cross-correlation of x(t) and y(t) is defined as

[0043]

[0044] The same applies Thus, the shaping operator f(t) is obtained (where N=m is the length of the filter operator)

[0045] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared with the prior art:

[0046] The method of the present invention eliminates the randomness of the wavelet extraction of the minimum offset or zero offset common center point gather through wavelet extraction, deconvolution processing and shaping operator application, and performs quality control by comparing the waveform consistency with the drilling and logging data to achieve well control fidelity processing. By suppressing the dynamic correction stretching distortion of the far-offset seismic data, the imaging distortion caused by the dynamic correction stretching is achieved, thereby improving the imaging quality and reliability of the seismic data and improving the utilization rate of the medium and shallow seismic data.

[0047] The target wavelet in the present invention is obtained through a series of seismic data preprocessing work such as pre-stack denoising, amplitude compensation, and deconvolution. The distortion caused by the interference waves during the propagation of the seismic wavelet and the energy and frequency attenuation in the propagation path are eliminated, and the original state of the seismic excited wavelet is restored as much as possible.

[0048] The present invention can be applied to the technical field of petroleum seismic exploration data analysis and processing, and is used for compressive correction of tensile distortion, improving the imaging quality of shallow-middle layer data, reducing the risk of drilling errors, and providing reliable data for efficient exploration and development of oil and gas fields.

[0049] The present invention will be described in further detail below with reference to the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a result diagram showing the comparison of the close-offset stacking data of the target wavelet extracted in an embodiment of the present invention and the waveform of the synthetic record of the drilling and logging data;

[0051] Figure 2 This is a comparison diagram of the gather results before and after compression braking and tensile distortion correction in an embodiment of the present invention;

[0052] Figure 3 This is a comparison diagram of the superimposed cross-section results before and after the compression and braking correction for tensile distortion in an embodiment of the present invention; DETAILED DESCRIPTION

[0053] Embodiment A method for correcting tensile distortion by pressing

[0054] This embodiment performs compression correction for tensile distortion on seismic data from a region in the Sichuan Basin. The specific method includes the following steps performed in sequence:

[0055] S1. Pre-stack preprocessing and common center point domain data rearrangement

[0056] Obtain and input the original single-shot seismic data after static correction, and obtain pre-processed data through pre-stack noise suppression, amplitude compensation and deconvolution processing;

[0057] Rearrange the pre-processed data into the common center point domain data to obtain the common center point gather data;

[0058] S2. Near and Far Offset Data Processing

[0059] In the offset data of the shot check point, the close offset data in the common center point gather data are screened out, and the close track stacking data are obtained through dynamic correction and stacking processing.

[0060] Compare the waveform consistency between the near-track stacking data and the drilling and logging synthetic records: Figure 1 As shown in the figure, the drilling and logging synthetic record has a high degree of agreement with the waveform of the near-offset stacked section passing through the well, and the frequency and phase consistency are high, indicating that the selected near-offset data in this area is appropriate. The near-offset stacked data is output to prepare data for target wavelet extraction.

[0061] If the waveform fit comparison result is not high, and the drilling and logging synthetic record and the seismic near-offset stacked section have low consistency in waveform, frequency, and phase, repeat the above steps and reselect appropriate near-offset data;

[0062] In the shot check point offset data, the far-offset data that has been stretched by dynamic correction in the common center point gather data is selected. After dynamic correction and stacking processing, the far-offset stacked data is output to prepare the data for the extraction of the wavelet to be shaped.

[0063] Among them, in the waveform consistency comparison, the VSP corridor stacking results can also be used instead of the drilling and logging synthetic records.

[0064] S3. Wavelet Extraction

[0065] Extract the near-track autocorrelation wavelet from the close-offset stacked data with high consistency as the target wavelet s1(t);

[0066] Extract the far-path autocorrelation wavelet from the far-offset stacked data that has generated the NMO stretch distortion as the wavelet to be shaped s2(t);

[0067] Among them, the far-offset data that has been corrected for stretch distortion Where Δf is the difference between the main frequency of the common-center point gather at far offset and the main frequency of the common-center point gather at close offset for the same reflection event, and f is the main frequency of the common-center point gather at close offset for the same reflection event.

[0068] S4. Deconvolution

[0069] Perform deconvolution on the target wavelet s1(t) and the wavelet to be shaped s2(t) to obtain the filter operator f(t);

[0070] S5. Apply the filter operator f(t) to the pre-stack gathers through convolution filtering to correct the tensile distortion. The result is as follows: Figure 2 , before the NMO stretching distortion, the gathers have obvious NMO stretching distortion in the far offset segment, as shown by the arrows in the figure, the event axis becomes thicker and the frequency decreases;

[0071] After the NMO stretching distortion is corrected, the NMO stretching distortion of the far-offset gather is well suppressed. As shown by the arrows in the figure, the events at near and far offsets are of the same thickness and have similar frequencies.

[0072] By comparing the superimposed cross-section effects obtained before and after applying the method of the present invention, the results are as follows: Figure 3 As shown in the figure, after the dynamic correction stretching distortion is suppressed, the continuity of the stacked profile event is improved, the data signal-to-noise ratio is increased, and the frequency of the shallow layer in the data is improved; as shown by the arrows in the figure, after the dynamic correction stretching distortion is suppressed by the method of the present invention, the stacked profile event becomes thinner and the continuity is increased, indicating that the method of the present invention can suppress the imaging distortion caused by dynamic correction stretching, thereby improving the imaging quality and reliability of seismic data.

[0073] It should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art may still modify the technical solutions described in the above embodiments or replace some of the technical features therein with equivalents. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A method for correcting tensile distortion by compression, characterized in that: It performs pre-stack data preprocessing and common center point domain data rearrangement on seismic data, and then uses well-seismic combination to process near and far offset data to extract the target wavelet s1(t) and the wavelet to be shaped s2(t); Perform deconvolution on the target wavelet s1(t) and the wavelet to be shaped s2(t) to obtain the filtering operator f(t); Apply the filter operator f(t) to the pre-stack gathers through convolution processing to correct the tensile distortion; The deconvolution process is as follows: 1) Determine the linear equations for the filter operator f(t) Assuming that the residual between the output f(t)*s2(t) of the filter operator f(t) and the target wavelet s1(t) is e(t), then e(t)=f(t)*s2(t)-s1(t); In actual calculations, the filter operator gradually approaches 0 as time t increases infinitely. The filter operator has finite energy and is a realizable filter operator. According to the least squares criterion, let the error function Solve the filter operator f(t), using the necessary conditions of the limit We get a linear equation system about f(t), where N and m are the lengths of the filter operator: 2) Obtain the filter operator f(t) by solving the overdetermined equation First, solve the overdetermined equation R T RA=R T X, where Target wavelet matrix Filter operator matrix available, According to the definition of autocorrelation: x(t) is a measure of similarity at different times in a time series; Expressed as follows but By analogy, we can conclude that: At the same time, the cross-correlation of x(t) and y(t) is defined as The same applies Thus, the shaping operator f(t) is obtained, where N=m is the length of the filtering operator.

2. The method for correcting tensile distortion by compression according to claim 1, characterized in that: The pre-stack data pre-processing comprises: acquiring seismic data, and performing pre-stack denoising, amplitude compensation and deconvolution processing to obtain pre-processed data; The common center point domain data rearrangement is: performing common center point domain data rearrangement on the pre-processed data to obtain common center point gather data.

3. The method for correcting tensile distortion by compression according to claim 2, characterized in that: In the near-offset and far-offset data processing, the method for near-offset data processing is: From the shot check point offset data, the near-offset data in the common center point gather data is selected. After dynamic correction and stacking processing, the near-track stacking data is obtained. The waveform consistency is then compared with the drilling and logging synthetic record or VSP corridor stacking result: If the waveform fit comparison result is high, the near-offset stack data is output, and the near-track autocorrelation wavelet is extracted from the near-offset stack data as the target wavelet s1(t); if the waveform fit comparison result is low, data screening is performed again.

4. The method for correcting tensile distortion by compression braking according to claim 2, characterized in that: In the near-far offset data processing, the far offset data processing method is as follows: From the offset data of the shot check point, the far-offset data with stretched distortion corrected by dynamic movement is selected from the common center point gather data. After dynamic movement correction and stacking, the far-offset stacked data is output. The far-path autocorrelation wavelet is extracted from the far-offset stacked data as the wavelet to be shaped s2(t).

5. The method for correcting tensile distortion by pressing according to claim 1 or 2, characterized in that: The deconvolution process is to solve the filter operator f(t) according to the least squares criterion and the overdetermined equation.

Images