Seismic velocity fusion method based on Poisson algorithm

Abstract

The invention discloses a seismic velocity fusion method based on a Poisson algorithm, which belongs to the technical field of seismic data processing and interpretation, and comprises the following steps: obtaining a superficial layer velocity model and a mid-deep layer velocity model; based on the near-surface tomography inversion ray density of the micro-logging constraint and the corresponding gather quality, determining a velocity fusion region of the shallow surface velocity model in depth; picking up a horizon from the obtained shallow surface layer velocity model and the obtained middle-deep layer velocity model; selecting the picked horizon as an interface, and respectively opening time windows upwards and downwards in the speed fusion area; based on the time window, calculating and fusing the speed model in the time window by using a Poisson algorithm, and outputting the fused speed model in the time window; and splicing the original speed model above the time window, the original speed model below the time window and the fused speed model in the time window to obtain a final shallow, medium and deep speed model. According to the method, the modeling precision can be improved.

Description

Technical Field

[0001] The present invention belongs to the field of petroleum seismic exploration technology, relates to the technical field of seismic data processing and interpretation, and specifically relates to a seismic velocity fusion method based on Poisson's algorithm. Background Art

[0002] With the huge demand for oil and gas resources due to the rapid development of the national economy, double-complex exploration areas with complex surface conditions and complex underground structures have gradually become an important field of oil and gas exploration. Due to the limitations of the prestack time migration method, for the migration and homing processing of complex area structures, especially high-steep structures or even complex structures with vertical inversion, obvious deviation of the structure position will occur during the homing process due to the drastic lateral velocity change. For complex area imaging, conducting prestack depth migration research can better determine the structure morphology.

[0003] The accuracy of the velocity model is the key to the success of depth migration. The tomographic inversion technology based on a single wave has become increasingly perfect. The surface model tomographic inversion is a non-linear model inversion technology that uses the travel time and path of seismic first arrival rays to invert the velocity structure of the medium. Wang Xiao proposed a multi-information constrained tomographic inversion technology, and Gao Lidong proposed a near-surface modeling technology based on constrained first arrival tomographic inversion. The reflected wave carries information about the underground interface and can reflect the structure and parameter distribution at finer scales at different depths and positions underground. Zhang Kai proposed a tomographic velocity inversion method for reflected waves under double-complex conditions. The first arrival tomographic inversion method can obtain a relatively accurate near-surface velocity within a certain depth range. In actual work areas, there are generally relatively rich micro-logging data. Using the micro-logging velocity stratification for constrained tomographic inversion can further improve the accuracy of the near-surface layer velocity model.

[0004] The original data used to establish the full velocity model is the near-surface velocity structure obtained by tomographic inversion based on the first arrival time, and the middle and deep layer velocity models established based on the prestack time migration profile. Due to the different seismic wave information used, the near-surface tomography and the middle and deep layer velocity models will show inconsistent overlapping of the velocity models at certain depths in the near-surface layer, and even a phenomenon of velocity model vacancy will occur at the depths between the end position of the tomographic velocity model and the start position of the middle and deep layer velocity models. Therefore, under the premise of considering the structural depth and lateral velocity change, relevant experts have carried out research on velocity fusion technology. Li Lianjun performed fusion processing on the velocity field by using two smoothing algorithms, namely Gaussian weighting and median filtering, on the updated velocity model.

[0005] Teng Houhua used the sliding window mean method and the inverse distance weighting method to fuse the shallow surface layer and the middle and deep layer velocity models.

[0006] The specific method is as follows: (1) Determine the credible region A of the shallow subsurface velocity model in depth according to the ray density of near-surface tomography inversion constrained by micro-logging, that is, the starting and ending ranges of fusion; (2) Use the velocity continuation reconstruction algorithm to correct the abnormal velocity values within region A; (3) Within the fusion region, use the sliding window mean method and the inverse distance weighting method to fuse the shallow subsurface and mid-deep subsurface velocity models; (4) Reverse the velocity model correction amount of the mid-deep subsurface, perform small smoothing on it to make it close to the tomography inversion surface, fill the irregular vacant part between the two surfaces using the velocity continuation reconstruction method, perform fusion using steps (2) and (3), and at the same time correct the seismic data to the matched small smoothed surface to carry out prestack depth migration velocity update. Summary of the Invention

[0007] In view of the deficiencies of the prior art, the present invention provides a seismic velocity fusion method based on the Poisson algorithm that can improve the modeling accuracy, thereby improving the imaging accuracy of seismic data.

[0008] The technical solution of the present invention is as follows: A seismic velocity fusion method based on the Poisson algorithm, comprising the following steps: Step 1, obtain the shallow subsurface velocity model and the mid-deep subsurface velocity model; Step 2, determine the velocity fusion region of the shallow subsurface velocity model in depth based on the ray density of near-surface tomography inversion constrained by micro-logging and the corresponding gather quality; pick horizons from the obtained shallow subsurface velocity model and mid-deep subsurface velocity model. Step 3, select the horizons picked in Step 2 as the interfaces, and open time windows upward and downward respectively within the velocity fusion region determined in Step 2. Step 4, based on the time windows opened in Step 3, use the Poisson algorithm to calculate and fuse the velocity models within the time windows, and output the fused velocity model within the time windows. Step 5, splice the original velocity model above the time window, the original velocity model below the time window, and the fused velocity model within the time windows output in Step 4 to obtain the final shallow, mid, and deep velocity model.

[0009] In Step 1, the shallow subsurface velocity model is obtained by a tomographic inversion method. When obtaining the shallow subsurface velocity model by the tomographic inversion method, pick the mixed first arrival times including direct waves, refracted waves, and diffracted waves.

[0010] In Step 1, the mid-deep subsurface velocity model is obtained by using velocity analysis technology to convert the root-mean-square velocity picked from well velocity or prestack time migration into layer velocity.

[0011] In step 4, based on the time window opened in step 3, the Poisson algorithm is used to calculate and fuse the velocity model within the time window. The steps for outputting the fused velocity model within the time window are as follows: First, input the velocity model within the time window opened in step 3 to generate a discrete point file, then generate a data structure space, then calculate the vector field, then solve the Poisson equation, next extract the isosurface, and finally output the fused velocity model within the time window.

[0012] The beneficial effects of the present invention are as follows: Aiming at the problem of low accuracy in prestack depth migration velocity modeling for conventional processing of complex surface data, the present invention studies a modeling method for fusing the near-surface velocity model and the mid-deep layer reflection wave velocity model, combines the advantages of both, takes into account both the shallow surface layer and the mid-deep layer, and can improve the modeling accuracy. The Poisson reconstruction mathematical algorithm adopted by the present invention breaks through the traditional Gaussian weighted smoothing and median filtering algorithms, eliminates the process of local segmentation and local fitting result stitching, and is an implicit surface reconstruction method based on the global. This method has the advantages of global fitting and local fitting, and has strong robustness to velocity anomaly points, making the stitched surface smooth and delicate after the target model is reconstructed. Brief Description of the Drawings

[0013] Figure 1 is a schematic flow chart of the present invention; Figure 2 is a flow chart of the Poisson algorithm related to the present invention; Figure 3 is a diagram of the shallow surface layer velocity model related to the present invention; Figure 4 is a diagram of the mid-deep layer velocity model related to the present invention; Figure 5 is a diagram of the fused velocity model using the method of the present invention. Detailed Embodiments

[0014] The following further illustrates the present invention through specific embodiments, but is not limited thereto.

[0015] See Figure 1 , one of many embodiments of the present invention, a seismic velocity fusion method based on the Poisson algorithm, includes the following steps: Step 1, obtain the shallow surface layer velocity model and the mid-deep layer velocity model, as Figure 3 , Figure 4 shown; Step 2, based on the ray density of the near-surface tomography inversion constrained by micro-logging and the corresponding trace gather quality, determine the velocity fusion region of the shallow surface layer velocity model in depth; pick horizons from the obtained shallow surface layer velocity model and mid-deep layer velocity model; Step 3: Select the horizon picked up in Step 2 as the interface, and open time windows upward and downward respectively within the velocity fusion region determined in Step 2; Step 4: Based on the time windows opened in Step 3, use the Poisson algorithm to calculate and fuse the velocity model within the time windows, and output the fused velocity model within the time windows; Step 5: Stitch together the original velocity model above the time windows, the original velocity model below the time windows, and the fused velocity model within the time windows output in Step 4 to obtain the final shallow, medium, and deep velocity model, that is, the updated initial full velocity model, as Figure 5 shown.

[0016] In the aforementioned Step 1, the shallow surface velocity model is obtained based on the tomography inversion method. When obtaining the shallow surface velocity model based on the tomography inversion method, pick up the mixed first arrival times including direct waves, refracted waves, and diffracted waves.

[0017] In the aforementioned Step 1, the acquisition of the medium and deep velocity model is achieved by using velocity analysis technology to transform the root-mean-square velocity picked up from well velocity or prestack time migration into layer velocity.

[0018] As Figure 2 shown, in the aforementioned Step 4, based on the time windows opened in Step 3, using the Poisson algorithm to calculate and fuse the velocity model within the time windows, and outputting the fused velocity model within the time windows includes the following steps: First, input the velocity model within the time windows opened in Step 3 to generate a discrete point file, then generate a data structure space, then calculate the vector field, then solve the Poisson equation, next extract the isosurface, and finally output the fused velocity model within the time windows.

Claims

1. A seismic velocity fusion method based on Poisson algorithm, characterized in that: The following steps are involved: Step 1, obtaining a shallow layer velocity model and a medium-deep layer velocity model; Step 2, based on the near-surface tomographic inversion ray density and the corresponding gather quality constrained by micro-logging, determine the velocity fusion area of ​​the shallow surface velocity model at depth; pick up the layer from the obtained shallow surface velocity model and the medium-deep velocity model; Step 3, select the layer picked up in step 2 as the interface, and open time windows upward and downward respectively in the velocity fusion area determined in step 2; Step 4: Based on the time window opened in step 3, the velocity model in the time window is calculated and fused using the Poisson algorithm, and the fused velocity model in the time window is output; Step 5: splice the original velocity model above the time window, the original velocity model below the time window, and the fused velocity model within the time window output in step 4 to obtain the final shallow, medium and deep velocity model.

2. The seismic velocity fusion method based on Poisson algorithm according to claim 1 is characterized in that: In the step 1, the shallow layer velocity model is obtained by a tomographic inversion method.

3. The seismic velocity fusion method based on Poisson algorithm according to claim 2 is characterized in that: When the shallow surface velocity model is obtained based on the tomographic inversion method, the mixed first arrival time of the direct wave, the mixed first arrival time of the refracted wave, and the mixed first arrival time of the return wave are picked up.

4. The seismic velocity fusion method based on Poisson algorithm according to claim 1 is characterized in that: In step 1, the mid-deep layer velocity model is obtained by converting the well velocity or the root mean square velocity picked up by prestack time migration into the layer velocity using velocity analysis technology.

5. The seismic velocity fusion method based on Poisson algorithm according to claim 1 is characterized in that: In the step 4, based on the time window opened in step 3, the velocity model in the time window is calculated and fused using the Poisson algorithm, and the fused velocity model in the time window is output, which includes the following steps: firstly, the velocity model in the time window opened in step 3 is input, a discrete point file is generated, then a data structure space is generated, then a vector field is calculated, then the Poisson equation is solved, then an isosurface is extracted, and finally the fused velocity model in the time window is output.

Images