阅读说明：本技术 一种地震数据初至波的外推方法 (Extrapolation method of seismic data first-motion wave ) 是由 许卓 张凤蛟 韩立国 于 2020-06-05 设计创作，主要内容包括：本发明专利提出了一种地震数据初至波的外推方法。首先,在炮集的初至波周围设置时间窗；其次,将参考道和目标道的波场在炮域进行反褶积和叠加,形成虚拟波场；再次,将参考道波场与虚拟波场在检波点域进行褶积和叠加,得到超级虚拟初至波；从次,判断是否采用循环迭代；除此之外,进行初至波外推,目标道位于目标炮集的检波器排列之外；然后,依次在炮域和检波点域进行反褶积叠加及褶积叠加得到外推的超级虚拟初至波；之后,判断是否采用循环迭代。本发明能有效增强初至波能量,提高初至波信噪比,同时通过初至波外推大大增加可用初至波的数据量,为后续地震数据处理提供更丰富的信息,并能很好地用于复杂近地表条件下采集的地震数据。(The invention provides an extrapolation method of seismic data first-motion waves. Firstly, setting a time window around a first arrival wave of a shot gather; secondly, deconvoluting and superposing wave fields of the reference track and the target track in a shot domain to form a virtual wave field; thirdly, performing convolution and superposition on the reference channel wave field and the virtual wave field in a detection point domain to obtain a super virtual first arrival wave; judging whether to adopt loop iteration or not; in addition, performing first arrival wave extrapolation, wherein the target track is positioned outside the detector arrangement of the target shot set; then, carrying out deconvolution superposition and convolution superposition in the shot domain and the detection point domain in sequence to obtain an extrapolated super virtual first arrival wave; then, whether loop iteration is adopted is judged. The method can effectively enhance the energy of the first-arrival waves, improve the signal-to-noise ratio of the first-arrival waves, greatly increase the data volume of the available first-arrival waves by extrapolation of the first-arrival waves, provide richer information for subsequent seismic data processing, and can be well used for seismic data acquired under complex near-surface conditions.)

1. A method for seismic data first arrival extrapolation, the method comprising:

(1) setting a time window around the first arrival wave of the original shot set, and intercepting a wave field containing the first arrival wave;

(2) for a target shot with a seismic source located at X, a reference track G (A | X) and a target track G (B | X) are taken, wherein G (A | X) represents a wave field which is located at X and recorded at the detector A, and in the Fourier domain, an amplitude term is omitted and can be expressed asWhere ω is the angular frequency, τ_XARepresenting a first arrival travel, and similarly, G (B | X) represents the wavefield with the source located at X and recorded at detector B,

(3) within the shot, the waveDeconvoluting the field G (A | X) and the wavefield G (B | X), and adding the deconvolution results at each stationary phase source to form a virtual wavefield G (B | A) between detectors A and B^virUnder the far field approximate condition, the calculation formula is as follows:

wherein k is the wave number, N_sIs the number of stationary phase sources for detector pairs a and B, representing the complex conjugate, as a regularization parameter;

(4) in the demodulator probe domain, the wave field G (A | X) and the virtual wave field G (B | A)^virPerforming convolution, and superposing convolution results at each stationary phase detector to obtain a super virtual first-arrival wave reconstructed by the seismic source at the X position and the detector B, wherein the calculation formula is as follows:

wherein N is_gIs the number of stationary phase detection points;

(5) determining whether to adopt loop iteration processing according to the quality of the super virtual first arrival wave, if the first arrival wave is still difficult to pick up, using the output of the step (4) as a new input, and repeating the step (3) and the step (4) until the signal-to-noise ratio of the super first arrival wave can be accepted;

(6) repeating the steps (2) to (5) for all the trace sets in the detector arrangement corresponding to the target shot set, and reconstructing a super virtual first arrival wave of the target shot set;

(7) extrapolating a first arrival wave of a target shot gather with a seismic source located at X, selecting a certain shot with the seismic source located at Y, and taking a reference track G (A | Y) and a target track G (C | Y), wherein the detector position A of G (A | Y) is located in the detector arrangement of the target shot gather, and the detector position C of G (C | Y) is located outside the detector arrangement of the target shot gather;

(8) deconvoluting the wave field G (A | Y) and the wave field G (C | Y) in the shot domain, and superposing deconvolution results at each stationary phase seismic source to form a space between the detector A and the detector CVirtual wave field G (C | A)^virUnder the far approximation condition, the calculation formula is as follows:

wherein N is_sIs the number of stationary phase sources for detector pairs a and C;

(9) in the demodulator probe domain, the wave field G (A | X) and the virtual wave field G (C | A)^virPerforming convolution calculation, and superposing convolution results at each stationary phase detector to obtain an extrapolated super virtual first-arrival wave reconstructed by the seismic source at the X position and the detector C, wherein the calculation formula is as follows:

(10) determining whether to adopt loop iteration processing according to the quality of the extrapolated super virtual first arrival wave, if the first arrival wave is still difficult to pick up, using the output of the step (9) as a new input, and repeating the step (8) and the step (9) until the signal-to-noise ratio of the super first arrival wave can be accepted;

(11) repeating the steps (7) to (10) for all the trace sets which can be extrapolated except the detector arrangement corresponding to the target shot set, and extrapolating the super virtual first arrival wave of the target shot set;

(12) and (3) combining the results of the step (6) and the step (11) for the target shot set, and adding the super virtual first-arrival wave extrapolated in the step (11) to the far offset direction based on the super virtual first-arrival wave reconstructed in the step (6) to form the super virtual first-arrival wave after the target shot is reconstructed and extrapolated.

2. The method for extrapolation of seismic data first arrival waves according to claim 1, wherein the time window of step (1) has a length of two to three periods including the expected first arrival.

3. The method for automatically picking up seismic wave first arrivals as claimed in claim 1, wherein the regularization operator in step (3) and step (8) is determined by: and taking different percentages of the average power spectrum of the target shot gather as regularization operators to obtain the super virtual first-break waves, and then using the regularization operator with the best obtaining effect for calculating the super virtual first-break waves of all the shot gathers.

Technical Field

The invention relates to the technical field of geophysical exploration seismic data processing, in particular to a seismic data first-arrival wave extrapolation method with strong noise immunity.

Background

The seismic wave velocity parameter is an extremely important physical property parameter in seismic exploration, plays an important role in each link of the seismic exploration, not only provides a main basis for processing and explaining seismic data, but also can indirectly reflect the structure and lithology information of an underground medium. Therefore, how to accurately obtain the velocity information of the underground medium is a hot problem of continuous and intensive research in the field of seismic exploration.

The first-arrival travel-time chromatographic inversion is a high-stability, accurate and efficient speed modeling method, based on a ray theory, the travel-time information of the seismic first-arrival waves is used for inverting the speed and the interface morphology of an underground medium, and the inversion result can be used for analyzing different lithologic interfaces, displaying the position and scale of an underground abnormal body, identifying a steep sill, a fracture structure and the like. However, the accuracy of the first-arrival travel time tomographic inversion has a direct relationship with the first-arrival pickup quality and the ray density, and the first-arrival pickup has a close relationship with the signal-to-noise ratio of seismic data, waveform variation among seismic channels, near-surface conditions and the like, so that under a complex terrain condition, the conventional first-arrival pickup method is difficult to obtain high-accuracy first-arrival pickup results and high-quality first-arrival information with abundant numbers.

The complex near-surface problem is always a big problem faced by seismic exploration, for example, in areas such as mountainous regions, deserts, Gobi, mountain front zones, loess tablelands and the like, the problems of complex surface conditions, severe topographic relief, large lateral difference of near-surface velocity and the like exist, so that the static correction problem is very prominent, and if the processing is not good, the results of subsequent seismic data processing such as offset imaging and the like are seriously affected. However, the conventional static correction method is difficult to meet the high-precision static correction requirement in the region with complex surface conditions. Compared with the traditional static correction method, the first-arrival travel-time chromatographic inversion technology is not restricted by changes of the earth surface and the near-earth surface structure, an accurate near-earth surface velocity model can be constructed, the static correction problem can be well solved, and the method has obvious advantages.

Therefore, with the continuous improvement of the seismic exploration precision requirement, in order to better solve the problem of complex surface seismic exploration, a technical means capable of improving the quality of first-arrival waves and increasing effective information of the first-arrival waves is needed, and the characteristic that chromatographic inversion needs as much data volume as possible during first-arrival travel is better met, so that a more precise velocity model closer to the actual underground condition can be established during first-arrival travel chromatographic inversion, and high-precision velocity information is provided for seismic exploration data processing and interpretation.

Disclosure of Invention

The invention aims to provide an effective technical means for improving the signal-to-noise ratio of first-arrival waves and extrapolating the first-arrival waves, which can solve the problem of low first-arrival picking precision under a complex surface condition on one hand, and provide richer information for chromatographic inversion during first-arrival travel through the extrapolated first-arrival waves on the other hand, so that a more detailed speed model is established.

To achieve the above object of the present invention, the following steps are taken:

(1) setting a time window around the first arrival wave of the original shot set, and intercepting a wave field containing the first arrival wave;

wherein the time window length is two to three cycles including the expected first arrival.

(2) For a target shot with a seismic source located at X, a reference track G (A | X) and a target track G (B | X) are taken, wherein G (A | X) represents a wave field which is located at X and recorded at the detector A, and in the Fourier domain, an amplitude term is omitted and can be expressed asWhere ω is the angular frequency, τ_XARepresenting a first arrival travel, and similarly, G (B | X) represents the wavefield with the source located at X and recorded at detector B,

(3) deconvoluting the wavefield G (A | X) and the wavefield G (B | X) in the shot domain, and superposing deconvolution results at each stationary phase seismic source to form a virtual wavefield G (B | A) between the detectors A and B^virUnder the far field approximate condition, the calculation formula is as follows:

wherein k is the wave number, N_sIs the number of stationary phase sources for detector pairs a and B, representing the complex conjugate, as a regularization parameter;

where the regularization operator typically takes a very small value, which may be a percentage of the average spectral power, e.g., 1% of the average spectral power.

(4) In the demodulator probe domain, the wave field G (A | X) and the virtual wave field G (B | A)^virPerforming convolution, and superposing convolution results at each stationary phase detector to obtain a super virtual first-arrival wave reconstructed by the seismic source at the X position and the detector B, wherein the calculation formula is as follows:

wherein N is_gIs the number of stationary phase detection points;

(5) determining whether to adopt loop iteration processing according to the quality of the super virtual first arrival wave, if the first arrival wave is still difficult to pick up, using the output of the step (4) as a new input, and repeating the step (3) and the step (4) until the signal-to-noise ratio of the super first arrival wave can be accepted;

(6) repeating the steps (2) to (5) for all the trace sets in the detector arrangement corresponding to the target shot set, and reconstructing a super virtual first arrival wave of the target shot set;

(7) extrapolating a first arrival wave of a target shot gather with a seismic source located at X, selecting a certain shot with the seismic source located at Y, and taking a reference track G (A | Y) and a target track G (C | Y), wherein the detector position A of G (A | Y) is located in the detector arrangement of the target shot gather, and the detector position C of G (C | Y) is located outside the detector arrangement of the target shot gather;

(8) deconvoluting the wave field G (A | Y) and the wave field G (C | Y) in the shot domain, and superposing deconvolution results at each stationary phase seismic source to form a virtual wave field G (C | A) between the detector A and the detector C^virUnder the far approximation condition, the calculation formula is as follows:

wherein N is_sIs the number of stationary phase sources for detector pairs a and C;

where the regularization operator typically takes a very small value, which may be a percentage of the average spectral power, e.g., 1% of the average spectral power.

(9) In the demodulator probe domain, the wave field G (A | X) and the virtual wave field G (C | A)^virPerforming convolution calculation, and superposing convolution results at each phase-stabilized detector to obtain an extrapolated super-virtual first-arrival wave reconstructed by the source at the X position and the detector C, wherein the calculation formula is as follows:

(10) determining whether to adopt loop iteration processing according to the quality of the extrapolated super virtual first arrival wave, if the first arrival wave is still difficult to pick up, using the output of the step (9) as a new input, and repeating the step (8) and the step (9) until the signal-to-noise ratio of the super first arrival wave can be accepted;

(11) repeating the steps (7) to (10) for all the trace sets which can be extrapolated except the detector arrangement corresponding to the target shot set, and extrapolating the super virtual first arrival wave of the target shot set;

(12) and (3) combining the results of the step (6) and the step (11) for the target shot set, and adding the super virtual first-arrival wave extrapolated in the step (11) to the far offset direction based on the super virtual first-arrival wave reconstructed in the step (6) to form the super virtual first-arrival wave after the target shot is reconstructed and extrapolated.

The regularization operator is determined by the following method: and taking different percentages of the average power spectrum of the target shot gather as regularization operators to obtain the super virtual first-break waves, and then using the regularization operator with the best obtaining effect for calculating the super virtual first-break waves of all the shot gathers.

The invention has the beneficial effects that: the energy of the first-motion waves can be effectively enhanced, the signal-to-noise ratio of the first-motion waves is improved, the number of the easily-identified first-motion waves is greatly increased, the available data of the first-motion waves are further expanded through first-motion wave extrapolation, the first-motion waves which are higher in quality and easier to pick can be provided for subsequent first-motion picking, the data volume of the first-motion waves picked during travel can be increased sharply, and therefore powerful bases are provided for static correction and chromatography inversion during first-motion travel under the complex near-surface condition, the method has good technical value, and in practical application, the method is universal.

Drawings

FIG. 1 is a complex relief surface geological model diagram of an embodiment of a method for extrapolation of first-arrival waves of seismic data according to the invention.

FIG. 2 is a diagram of low SNR seismic single shot data according to an embodiment of the present invention.

FIG. 3 is a reconstructed and extrapolated result of a super-virtual first-arrival extrapolation of a seismic data first-arrival extrapolation method according to an embodiment of the present invention.

Detailed Description

The invention will be described in more detail below with reference to the accompanying drawings.

A method for extrapolating first-arrival waves of seismic data, recorded under the complex relief surface geological model shown in fig. 1, to the low signal-to-noise ratio seismic shot gather data of fig. 2, comprising the steps of:

(1) and setting a time window around the first arrival wave of the original shot set, and intercepting a wave field containing the first arrival wave.

The time window length is two to three cycles including the expected first arrival.

(2) For a target shot with a seismic source located at X, a reference track G (A | X) and a target track G (B | X) are taken, wherein G (A | X) represents a wave field which is located at X and recorded at the detector A, and in the Fourier domain, an amplitude term is omitted and can be expressed asWhere ω is the angular frequency, τ_XARepresenting a first arrival travel, and similarly, G (B | X) represents the wavefield with the source located at X and recorded at detector B,

(3) deconvoluting the wavefield G (A | X) and the wavefield G (B | X) in the shot domain, and superposing deconvolution results at each stationary phase seismic source to form a virtual wavefield G (B | A) between the detectors A and B^virUnder the far field approximate condition, the calculation formula is as follows:

wherein k is the wave number, N_sIs the number of stationary phase sources for detector pairs a and B, which represent the complex conjugate, which is the regularization parameter.

Where the regularization operator typically takes a very small value, which may be a percentage of the average spectral power, e.g., 1% of the average spectral power.

(4) In the demodulator probe domain, the wave field G (A | X) and the virtual wave field G (B | A)^virPerforming convolution, and superposing convolution results at each stationary phase detector to obtain a super virtual first-arrival wave reconstructed by the seismic source at the X position and the detector B, wherein the calculation formula is as follows:

wherein N is_gIs the number of stationary phase detection points.

(5) And (3) determining whether to adopt loop iteration processing according to the quality of the super virtual first arrival wave, if the first arrival wave is still difficult to pick up, using the output of the step (4) as a new input, and repeating the step (3) and the step (4) until the signal-to-noise ratio of the super first arrival wave is acceptable.

(6) And (5) repeating the steps (2) to (5) for all the trace sets in the detector arrangement corresponding to the target shot set, and reconstructing the super virtual first-arrival wave of the target shot set.

(7) And extrapolating the first arrival wave of a target shot gather with the seismic source located at X, selecting a certain shot with the seismic source located at Y, and taking a reference track G (A | Y) and a target track G (C | Y), wherein the detector position A of G (A | Y) is located in the detector array of the target shot gather, and the detector position C of G (C | Y) is located outside the detector array of the target shot gather.

(8) Deconvoluting the wave field G (A | Y) and the wave field G (C | Y) in the shot domain, and superposing deconvolution results at each stationary phase seismic source to form a virtual wave field G (C | A) between the detector A and the detector C^virComputing the metric under the condition of distant approximationThe formula is as follows:

wherein N is_sIs the number of stationary phase sources for detector pairs a and C.

Where the regularization operator typically takes a very small value, which may be a percentage of the average spectral power, e.g., 1% of the average spectral power.

(9) In the demodulator probe domain, the wave field G (A | X) and the virtual wave field G (C | A)^virPerforming convolution calculation, and superposing convolution results at each phase-stabilized detector to obtain an extrapolated super-virtual first-arrival wave reconstructed by the source at the X position and the detector C, wherein the calculation formula is as follows:

(10) and (3) determining whether to adopt loop iteration processing according to the quality of the extrapolated super virtual first arrival wave, if the first arrival wave is still difficult to pick up, using the output of the step (9) as a new input, and repeating the step (8) and the step (9) until the signal-to-noise ratio of the super first arrival wave is acceptable.

(11) And (5) repeating the steps (7) to (10) for all the trace sets which can be extrapolated except the detector arrangement corresponding to the target shot set, and extrapolating the super virtual first arrival wave of the target shot set.

(12) And (3) combining the results of the step (6) and the step (11) for the target shot set, and adding the super virtual first-arrival wave extrapolated in the step (11) to the far offset direction based on the super virtual first-arrival wave reconstructed in the step (6) to form the super virtual first-arrival wave after the target shot is reconstructed and extrapolated.

The regularization operator is determined by the following method: and taking different percentages of the average power spectrum of the target shot gather as regularization operators to obtain the super virtual first-break waves, and then using the regularization operator with the best obtaining effect for calculating the super virtual first-break waves of all the shot gathers.

FIG. 3 is a reconstructed and extrapolated super virtual first-arrival result obtained using the techniques of the present invention. It can be seen that, compared with the original low signal-to-noise ratio seismic single shot data (fig. 2), the method for extrapolating the first-arrival waves of seismic data disclosed by the invention can effectively extrapolate the first-arrival waves while improving the signal-to-noise ratio of the first-arrival waves, so that the data volume of available first-arrival waves is greatly increased.

Various modifications and alterations to the embodiments described above will become apparent to those skilled in the art, and the embodiments of the invention are not to be considered limiting of the invention, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein and any such modifications, equivalents, improvements and equivalents as are within the scope of the invention.