Wave equation waveform inversion processing method, device, equipment and medium

Abstract

The invention relates to a wave equation waveform inversion processing method and device, equipment and a medium. The processing method comprises the following steps: establishing an initial speed model; determining a corresponding relation among the full wave field data, the transmitted wave data and the reflected wave data, and splitting the original seismic shot gather record into predicted transmitted wave data and predicted reflected wave data; constructing a target functional by minimizing the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data under the meaning of least squares; according to the target functional, utilizing a model parameter gradient calculation formula to construct an inversion parameter gradient; and carrying out iterative updating to obtain a speed model after iterative updating. Therefore, based on waveform inversion of the wave equation combined with the transmitted wave and the reflected wave, the reflected wave has a deep illumination effect and can be used for recovering medium and low wave number components of the deep part in the model and overcoming the problem of too little long offset data to update the background model of the deep part, and the inversion accuracy is improved.

Description

Technical Field

[0001] The present application relates to the technical field of oil exploration seismic data preprocessing, and in particular to a wave equation waveform inversion processing method, device, equipment and medium for combining transmission wave and reflection wave. Background Art

[0002] Full Waveform Inversion (FWI) is a wave equation inversion method based on waveform matching. In theory, FWI can recover all model wave number components from low to high in the seismic frequency band. However, full waveform inversion has high requirements for observation data, and high-density, all-round seismic waveform data is required to obtain a more accurate underground medium model. If the data quality is not high or the data coverage is incomplete, the inversion results may be affected. Full waveform inversion is sensitive to the selection of the initial model, and an unreasonable initial model may lead to unstable or inaccurate inversion results.

[0003] Due to the periodicity of seismic wave signals, the objective function of waveform matching has many local extreme values, that is, when the distance between the simulated data and the observed data is greater than half a cycle, a cycle jump will occur. This requires that the inversion must have a good initial model or contain low-frequency and long offset data. Conventional FWI usually uses transmission wave data to restore the medium and low wavenumber characteristics of the model, and then uses small-angle reflection wave data to update the high wavenumber components of the model. However, this is limited by the penetration depth of the transmission wave, making it impossible for conventional FWI to update the deep background model. For the update of the deep background model, long offset data is crucial, but it is difficult to obtain enough long offset data to update the deep medium and low wavenumber components in actual acquisition. Therefore, how to improve the inversion accuracy and overcome the problem of insufficient long offset data to update the deep background model needs to be solved urgently. Summary of the invention

[0004] Based on this, it is necessary to provide a wave equation waveform inversion processing method, device, equipment and medium that can overcome the problem of insufficient long offset data to update the deep background model and improve the inversion accuracy.

[0005] In a first aspect, the present application provides a wave equation waveform inversion processing method, comprising the following steps:

[0006] Establish initial velocity model;

[0007] According to the initial velocity model, the correspondence between the full wave field data, the transmission wave data and the reflection wave data is determined, and the original seismic shot gather records of the detected full wave field data are split into the predicted transmission wave data and the predicted reflection wave data according to the correspondence;

[0008] Construct an objective functional by minimizing the differences between the predicted and observed reflected wave data and between the predicted and observed transmitted wave data in the least - squares sense;

[0009] Construct the inversion parameter gradient according to the objective functional using the model parameter gradient calculation formula;

[0010] Iteratively update the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model.

[0011] In one embodiment, the correspondence between the full - wavefield data, transmitted wave data, and reflected wave data is determined based on the initial velocity model, and is determined using the following formula:

[0012] d(m 0 )=Ru(m 0 )=Ru 0 (m 0 )+Rδu(m 0 )

[0013] where m 0 is the initial velocity model, d is the full - wavefield data of the seismic data detected by the geophones, R represents the sampling operator of the geophone points, Ru 0 (m 0 ) represents the transmitted wave in the seismic data, and Rδu(m 0 ) represents the primary reflected wave in the seismic data.

[0014] In one embodiment, in constructing the objective functional by minimizing the differences between the predicted and observed reflected wave data and between the predicted and observed transmitted wave data in the least - squares sense, the objective functional formula C(m 0 ) is as follows:

[0015]

[0016] Δd t =Ru 0 -d t

[0017] Δd r =Rδu - d r

[0018] where Δd t is the data residual of the transmitted wave, Δd r is the data residual of the reflected wave, Ru 0 is the predicted transmitted wave data, d t is the observed transmitted wave data, Rδu is the predicted reflected wave data, d r is the observed reflected wave data, and γ is the damping coefficient.

[0019] In one embodiment, in constructing the inversion parameter gradient according to the objective functional by using the model parameter gradient calculation formula, the model parameter gradient calculation formula is as follows:

[0020]

[0021] Wherein, is the adjoint wave field, is the adjoint perturbation wave field.

[0022] In one embodiment, after iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model, the wave equation waveform inversion processing method further includes the following steps:

[0023] Output a low-to-medium wavenumber velocity field according to the iteratively updated initial velocity model.

[0024] In a second aspect, the present application provides a wave equation waveform inversion processing device, and the device includes:

[0025] An initial velocity model establishment module, configured to establish an initial velocity model;

[0026] A correspondence determination and splitting module, configured to determine the correspondence relationships of the full wave field data, the transmitted wave data, and the reflected wave data according to the initial velocity model, and split the original seismic shot gather record of the detected full wave field data into the predicted transmitted wave data and the predicted reflected wave data according to the correspondence relationships;

[0027] An objective functional construction module, configured to construct an objective functional by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data in the least squares sense;

[0028] An inversion parameter gradient construction module, configured to construct an inversion parameter gradient according to the objective functional by using the model parameter gradient calculation formula;

[0029] An iterative update module, configured to iteratively update the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model.

[0030] In one embodiment, in the correspondence determination and splitting module, in determining the correspondence relationships of the full wave field data, the transmitted wave data, and the reflected wave data according to the initial velocity model, the following formula is used for determination:

[0031] d(m 0 ) = Ru(m 0 ) = Ru 0 (m 0 ) + Rδu(m0 )

[0032] Among them, m 0 is the initial velocity model, d is the full-wavefield data of the seismic data detected by the geophone, R represents the sampling operator of the geophone point, and Ru 0 (m 0 ) represents the transmitted wave in the seismic data, and Rδu(m 0 ) represents the primary reflected wave in the seismic data.

[0033] In one embodiment, in the target functional construction module, in the least squares sense, the target functional is constructed by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data. The target functional formula C(m 0 ) is as follows:

[0034]

[0035] Δd t = Ru 0 - d t

[0036] Δd r = Rδu - d r

[0037] Among them, Δd t is the data residual of the transmitted wave, Δd r is the data residual of the reflected wave, Ru 0 is the predicted transmitted wave data, d t is the observed transmitted wave data, Rδu is the predicted reflected wave data, and d r is the observed reflected wave data, and γ is the damping coefficient.

[0038] In a third aspect, the present application provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the steps of the wave equation waveform inversion processing method described in any of the above embodiments.

[0039] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the wave equation waveform inversion processing method described in any of the above embodiments.

[0040] The wave equation waveform inversion processing method comprises establishing an initial velocity model; determining the correspondence between full wave field data, transmission wave data and reflection wave data according to the initial velocity model, and splitting the original seismic shot gather record of the detected full wave field data into predicted transmission wave data and predicted reflection wave data according to the correspondence; constructing a target functional by minimizing the difference between the predicted reflection wave data and the observed reflection wave data and the difference between the predicted transmission wave data and the observed transmission wave data in the least square sense; constructing an inversion parameter gradient according to the target functional using a model parameter gradient calculation formula; iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain an iteratively updated velocity model; thus, based on the wave equation waveform inversion combined with the transmission wave and the reflection wave, the reflection wave has a deep illumination effect, can be used to restore the medium and low wave number components in the deep part of the model, can overcome the problem of insufficient long offset data to update the deep background model, and improve the inversion accuracy. The above-mentioned wave equation waveform inversion processing method is based on the use of reflection wave information to update the deep part of the model. This technology combines the advantages of FWI and RWI (Reflection Wavelet Interferometry, reflection wave waveform inversion) inversion, establishes a joint inversion objective function of transmission wave and reflection wave to obtain accurate low-frequency background, and uses the low- and medium-wavenumber components of the shallow layer restored by the transmission wave waveform inversion, and uses the reflection wave waveform inversion to supplement the medium-wavenumber components of the medium- and deep-layer model. This application is particularly suitable for deep-water deep-ground velocity modeling problems. Starting from the transmission wave and reflection wave inversion, a transmission wave and reflection wave field joint waveform inversion based on high-precision wave theory is formed, forming a set of stable transmission wave and reflection wave joint full waveform inversion technology, and realizing high-precision velocity modeling of the shallow, medium and deep layers of the model. According to the difference in the wave paths of the transmission wave and the reflection wave, the present invention combines the corresponding path update direction, effectively makes up for the shortcomings of the traditional full waveform inversion, and realizes the simultaneous balanced modeling of the shallow, medium and deep layers of the model, which improves the inversion accuracy compared with the traditional method. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A schematic diagram of the steps of a wave equation waveform inversion processing method according to an embodiment;

[0042] Figure 2a is the full wave field data Ru(m 0 );

[0043] Figure 2b is the transmission wave data Ru in the seismic data 0 (m 0 );

[0044] Figure 2c is the first reflection wave data Rδu(m 0 );

[0045] Figure 3a is the velocity update amount of the wave equation inversion for the transmitted wave data in seismic data;

[0046] Figure 3b is the velocity update amount of the wave equation inversion for the primary reflected wave data in seismic data;

[0047] Figure 3c is Figure 3a and Figure 3b the velocity update amount obtained by weighting;

[0048] Figure 4 is the schematic diagram of the module structure of the wave equation waveform inversion processing device in an embodiment;

[0049] Figure 5 is the schematic diagram of the structure of a computer device in an embodiment. Detailed implementation manners

[0050] For the convenience of understanding the present application, in order to make the above objects, features, and advantages of the present application more obvious and understandable, the following will describe the detailed implementation manners of the present application with reference to the accompanying drawings. In the following description, many specific details are set forth to fully understand the present application, and the preferred embodiments of the present application are given in the accompanying drawings. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive. The present application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present application. Therefore, the present application is not limited by the specific embodiments disclosed below. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of the present application, the meaning of "a plurality" is at least two, such as two, three, etc., unless otherwise specifically defined. In the description of the present application, the meaning of "several" is at least one, such as one, two, etc., unless otherwise specifically defined. When an element is considered to be "connected" to another element, it may be directly connected to the other element or there may be intermediate elements present simultaneously. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art belonging to the technical field of the present application. The terms used herein are only for the purpose of describing specific implementation manners and are not intended to limit the present application. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0051] Embodiment 1

[0052] In a first aspect, the present application provides a method for processing wave equation waveform inversion. Please refer to Figure 1 , the method for processing wave equation waveform inversion includes the following steps:

[0053] S110: Establish an initial velocity model;

[0054] In this embodiment, the initial velocity model can be established based on existing technical means. It should be noted that in the wave equation, the initial velocity model refers to the velocity distribution in the wave field at the initial moment. The change of the initial velocity model will affect the propagation and evolution of seismic waves. Therefore, for seismic research, it is very important to understand the initial velocity model. By analyzing and studying the initial velocity model, the structure and properties of the underground medium can be better understood, providing a reliable basis for research in fields such as earthquake prediction, mineral resource exploration, and environmental geology. It should be noted that the initial velocity model is only an approximate solution because in actual seismic observations, the propagation and evolution of seismic waves are affected by many factors, such as the inhomogeneity of the underground medium, complex terrain and landforms, and the characteristics of seismic sources. Therefore, it is necessary to conduct comprehensive analysis and interpretation in combination with other geological data and technical means.

[0055] In this embodiment, through the initial velocity model, the initial velocity model is iteratively updated and optimized based on transmitted waves and reflected waves later.

[0056] S120: According to the initial velocity model, determine the correspondence between the full wave field data, transmitted wave data, and reflected wave data, and split the original seismic shot gather record of the detected full wave field data into predicted transmitted wave data and predicted reflected wave data according to the correspondence;

[0057] In this embodiment, according to the established initial velocity model, determine the correspondence between the full wave field data, transmitted wave data, and reflected wave data. In one embodiment, the correspondence between the full wave field data, transmitted wave data, and reflected wave data is determined according to the initial velocity model by using the following formula:

[0058] d(m 0 ) = Ru(m 0 ) = Ru 0 (m 0 ) + Rδu(m 0 )

[0059] where m 0 is the initial velocity model, d is the full wave field data of the seismic data detected by the geophone, R represents the sampling operator of the geophone point, Ru 0 (m 0 ) represents the transmitted wave in the seismic data, and Rδu(m 0) represents a reflected wave in seismic data. Thus, by using the above formula, the corresponding relationships of the full-wavefield data, transmitted wave data, and reflected wave data can be better determined based on the initial velocity model.

[0060] In this example, considering that the underground disturbance has a certain continuity and spatial distribution, the superposition of the scattered wavefields forms a reflected wave (the scattered wavefield mentioned below is the reflected wave). In seismic exploration, the seismic wavefield propagates to the surface and is collected by a limited series of geophones to form observed data (ignoring multiple scattering). For example, a reflected wave refers to the part of the wave that is reflected back into the original medium and continues to propagate when the wave encounters the interface of the medium. A transmitted wave is the part of the wave that passes through the interface and enters another medium and continues to propagate when the wave passes through the interface of the medium.

[0061] S130: In the least-squares sense, construct an objective functional by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data;

[0062] In this embodiment, the goal of seismic wave inversion is to find an appropriate solution in the model space to minimize the residual between the predicted data and the observed data. Based on the framework of the least-squares sense, the objective functional is defined. In one embodiment, in the construction of the objective functional by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data in the least-squares sense, the objective functional formula C(m 0 ) is as follows:

[0063]

[0064] Δd t = Ru 0 - d t

[0065] Δd r = Rδu - d r

[0066] where, Δd t is the data residual of the transmitted wave, Δd r is the data residual of the reflected wave, Ru 0 is the predicted transmitted wave data, d t is the observed transmitted wave data, Rδu is the predicted reflected wave data, d r is the observed reflected wave data, and γ is the damping coefficient. Thus, by using the above objective functional formula, in the least-squares sense, both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data can be minimized.

[0067] S140: Construct the inversion parameter gradient according to the objective functional using the model parameter gradient calculation formula.

[0068] In this embodiment, construct the inversion parameter gradient according to the constructed objective functional using the model parameter gradient calculation formula. The goal of inversion is to find the appropriate m 0 model to minimize the objective functional Cm 0 In one embodiment, in the step of constructing the inversion parameter gradient according to the objective functional using the model parameter gradient calculation formula, the model parameter gradient calculation formula is as follows:

[0069]

[0070] where is the adjoint wavefield, and is the adjoint perturbation wavefield. Comparing with traditional FWI, the key of RWI is that it can better extract the reflection wave path information, and there is no high-frequency interference in the gradient formula. Thus, a joint inversion framework of transmission wave and reflection wave of the wave equation can be better established.

[0071] S150: Iteratively update the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model.

[0072] In this embodiment, iteratively update the initial velocity model m 0 with the obtained inversion parameter gradient, and determine whether the iteratively updated velocity model satisfies the convergence condition. If it satisfies the convergence condition, obtain the iteratively updated velocity model; if it does not satisfy the convergence condition, continue to repeat the foregoing steps. In this embodiment, update the velocity model based on the inversion framework, thereby ensuring that the medium and low wavenumber velocity field can be further determined subsequently.

[0073] In one embodiment, after the step of iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model, the wave equation waveform inversion processing method further includes the following steps:

[0074] Output the medium and low wavenumber velocity field according to the iteratively updated initial velocity model.

[0075] Thus, the medium and low wavenumber components in the deep part of the model can be updated and restored according to the initial velocity model, and the medium wavenumber components in the middle and deep layers of the model can be supplemented by using the reflection wave waveform inversion, which can further improve the inversion accuracy.

[0076] The wave equation waveform inversion processing method comprises establishing an initial velocity model; determining the correspondence between full wave field data, transmission wave data and reflection wave data according to the initial velocity model, and splitting the original seismic shot gather record of the detected full wave field data into predicted transmission wave data and predicted reflection wave data according to the correspondence; constructing a target functional by minimizing the difference between the predicted reflection wave data and the observed reflection wave data and the difference between the predicted transmission wave data and the observed transmission wave data in the least square sense; constructing an inversion parameter gradient according to the target functional using a model parameter gradient calculation formula; iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain an iteratively updated velocity model; thus, based on the wave equation waveform inversion combined with the transmission wave and the reflection wave, the reflection wave has a deep illumination effect, can be used to restore the medium and low wave number components in the deep part of the model, can overcome the problem of insufficient long offset data to update the deep background model, and improve the inversion accuracy. The above-mentioned wave equation waveform inversion processing method is based on the use of reflection wave information to update the deep part of the model. This technology combines the advantages of FWI and RWI (Reflection Wavelet Interferometry, reflection wave waveform inversion) inversion, establishes a joint inversion objective function of transmission wave and reflection wave to obtain accurate low-frequency background, and uses the low- and medium-wavenumber components of the shallow layer restored by the transmission wave waveform inversion, and uses the reflection wave waveform inversion to supplement the medium-wavenumber components of the medium- and deep-layer model. This application is particularly suitable for deep-water deep-ground velocity modeling problems. Starting from the transmission wave and reflection wave inversion, a transmission wave and reflection wave field joint waveform inversion based on high-precision wave theory is formed, forming a set of stable transmission wave and reflection wave joint full waveform inversion technology, and realizing high-precision velocity modeling of the shallow, medium and deep layers of the model. According to the difference in the wave paths of the transmission wave and the reflection wave, the present invention combines the corresponding path update direction, effectively makes up for the shortcomings of the traditional full waveform inversion, and realizes the simultaneous balanced modeling of the shallow, medium and deep layers of the model, which improves the inversion accuracy compared with the traditional method.

[0077] Example 2

[0078] The present application is further described below in conjunction with a wave equation waveform inversion processing method according to a specific embodiment.

[0079] It should be noted that for the update of the deep background model, long-offset data is crucial. However, it is difficult to obtain sufficient long-offset data in actual acquisition to update the low and medium wavenumber components in the deep part. The reflected wave has a depth illumination effect and can be used to recover the low and medium wavenumber components in the deep part of the model. Therefore, how to use the reflected wave information for the update of the deep part of the model has increasingly become the focus of seismic inversion. Inspired by the MBTT method, Xu et al. (2012) proposed the RWI method to obtain the update of the low and medium wavenumber components in the model through the information of the reflected wave path. In recent years, due to the illumination depth of the reflected wave, the RWI method has gradually received extensive attention. It essentially obtains a high wavenumber reflectivity model through the LSRTM method, then uses reverse migration to obtain predicted data, and back-projects the residual between it and the observed data onto the reflected wave path for background velocity update. This technology combines the advantages of FWI and RWI inversion, establishes a joint inversion objective function of transmitted wave and reflected wave to obtain an accurate low-frequency background, uses the transmitted wave waveform inversion to recover the shallow low and medium wavenumber components, and uses the reflected wave waveform inversion to supplement the medium wavenumber components in the medium and deep layer models. The present invention mainly aims at the problem of deep water and deep earth velocity modeling. Starting from the inversion of transmitted wave and reflected wave respectively, it forms a joint waveform inversion of transmitted wave and reflected wave field based on high-precision wave theory, and forms a set of stable joint full waveform inversion technology of transmitted wave and reflected wave to achieve high-precision velocity modeling in the shallow, medium and deep parts of the model.

[0080] (1) Technical principle

[0081] When the underground disturbance has a certain continuity and spatial distribution, the superposition of the scattered wave fields forms a reflected wave (the scattered wave field mentioned below is the reflected wave). In seismic exploration, the seismic wave field propagates to the surface and is collected by a limited series of geophones to form observed data (ignoring multiple scattering):

[0082] d(m 0 ) = Ru(m 0 ) = Ru 0 (m 0 ) + Rδu(m 0 )(1)

[0083] In the formula, m 0 is the initial velocity model, d represents the full wave field data of the seismic record, and R represents the sampling operator of the geophone point. Ru 0 (m 0 ) and Rδu(m 0 ) respectively correspond to the direct wave and the primary reflected wave in the seismic data.

[0084] The goal of seismic wave inversion is to find an appropriate solution in the model space to minimize the residual between the predicted data and the observed data. In order to recover the low and medium wavenumber components of the model, in the framework of the least squares sense, the objective functional is defined as:

[0085]

[0086] where Δd t = Ru 0 - d t and Δd r = Rδu - d r are the data residuals of the transmitted wave and the reflected wave respectively, Ru 0 and d t represent the predicted transmitted wave data and the observed transmitted wave data respectively, and Rδu and d r represent the predicted reflected wave data and the observed reflected wave data respectively. The goal of inversion is to find a suitable m 0 model to minimize the objective functional Cm 0 .

[0087]

[0088] where are the adjoint wavefield and the adjoint perturbed wavefield respectively. Comparing with the traditional FWI, the key of RWI lies in that it can better extract the reflection wave path information, and there is no high-frequency interference in the gradient formula.

[0089] According to the differences between the transmitted wave and the reflected wave paths, the present invention combines the corresponding path update directions, effectively makes up for the deficiencies of the traditional full waveform inversion, and realizes the simultaneous balanced modeling of the shallow, middle and deep layers of the model.

[0090] The technical implementation steps are as follows:

[0091] Input the original seismic shot gather records (such as Figure 2a ) and the initial velocity model.

[0092] According to the shallow layer velocity of the model, as shown in equation (1), split the original seismic shot gather records into the transmitted wave ( Figure 2b ) and the reflected wave data ( Figure 2c );

[0093] Establish the joint inversion framework of the transmitted wave and the reflected wave of the wave equation according to equation (2);

[0094] Based on the inversion framework of steps (2) and (3), update the velocity model to obtain as Figures 2a to 2c shown.

[0095] Figure 2a is the full wavefield data Ru(m 0 ) of the seismic data, Figure 2b is the transmitted wave data Ru 0 (m 0 ) in the seismic data, Figure 2c is the primary reflected wave data Rδu(m in the seismic data0 )。

[0096] Figure 3a is the updated velocity obtained by inverse wave equation of transmitted wave data in seismic data according to the present application, corresponding to the first term on the right side of formula (2). Figure 3b is the updated velocity obtained by inverse wave equation of primary reflected wave data in seismic data according to the present application, corresponding to the second term on the right side of formula (2). Figure 3c is Figure 3a and Figure 3b weighted to obtain the updated velocity, corresponding to the weighted superposition of the two terms on the right side of formula (2).

[0097] The present invention mainly aims at the problem of velocity modeling in deep water and deep earth. According to the different inversion regions of transmitted waves and reflected waves, a joint waveform inversion of transmitted wave and reflected wave fields based on high-precision wave theory is formed, which solves the problems of low wavenumber in shallow layer and high wavenumber in deep layer in traditional full waveform inversion, forms a set of stable joint full waveform inversion technology of transmitted wave and reflected wave, and realizes the simultaneous balanced modeling of shallow, medium and deep layers of the model. The invention has broad application prospects in the fields of seismic exploration and prospecting.

[0098] Embodiment 3

[0099] Secondly, the present application provides a wave equation waveform inversion processing device. Please refer to Figure 4 , the device includes:

[0100] An initial velocity model establishment module, configured to establish an initial velocity model;

[0101] A correspondence determination and splitting module, configured to determine the correspondence between full wave field data, transmitted wave data and reflected wave data according to the initial velocity model;

[0102] A target functional construction module, configured to construct a target functional by minimizing the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data in the least squares sense;

[0103] An inversion parameter gradient construction module, configured to construct an inversion parameter gradient according to the target functional by using a model parameter gradient calculation formula;

[0104] An iterative update module, configured to perform iterative update on the initial velocity model with the obtained inversion parameter gradient to obtain an iteratively updated velocity model.

[0105] The wave equation waveform inversion processing device comprises the initial velocity model establishment module, the corresponding relationship determination and splitting module, the target functional construction module, the inversion parameter gradient construction module, and the iterative update module, and adopts the wave equation waveform inversion processing method, including establishing an initial velocity model; determining the corresponding relationship between the full wave field data, the transmission wave data and the reflection wave data according to the initial velocity model, and splitting the original seismic shot gather record of the detected full wave field data into predicted transmission wave data and predicted reflection wave data according to the corresponding relationship; in the sense of least squares, the difference between the predicted reflection wave data and the observed reflection wave data and the difference between the predicted transmission wave data and the observed transmission wave data are minimized to construct the target functional; according to the target functional, the inversion parameter gradient is constructed by using the model parameter gradient calculation formula; the initial velocity model is iteratively updated with the obtained inversion parameter gradient to obtain the iteratively updated velocity model; in this way, based on the wave equation waveform inversion of the combination of the transmission wave and the reflection wave, the reflection wave has a deep lighting effect, which can be used to restore the medium and low wave number components in the deep part of the model, and can overcome the problem of insufficient long offset data to update the background model in the deep part, thereby improving the inversion accuracy. The above-mentioned wave equation waveform inversion processing method is based on the use of reflection wave information to update the deep part of the model. This technology combines the advantages of FWI and RWI (Reflection Wavelet Interferometry, reflection wave waveform inversion) inversion, establishes a joint inversion objective function of transmission wave and reflection wave to obtain accurate low-frequency background, and uses the low- and medium-wavenumber components of the shallow layer restored by the transmission wave waveform inversion, and uses the reflection wave waveform inversion to supplement the medium-wavenumber components of the medium- and deep-layer model. This application is particularly suitable for deep-water deep-ground velocity modeling problems. Starting from the transmission wave and reflection wave inversion, a transmission wave and reflection wave field joint waveform inversion based on high-precision wave theory is formed, forming a set of stable transmission wave and reflection wave joint full waveform inversion technology, and realizing high-precision velocity modeling of the shallow, medium and deep layers of the model. According to the difference in the wave paths of the transmission wave and the reflection wave, the present invention combines the corresponding path update direction, effectively makes up for the shortcomings of the traditional full waveform inversion, and realizes the simultaneous balanced modeling of the shallow, medium and deep layers of the model, which improves the inversion accuracy compared with the traditional method.

[0106] In one embodiment, in the corresponding relationship determination and splitting module, the corresponding relationship between the full wave field data, the transmission wave data and the reflection wave data is determined according to the initial velocity model by using the following formula:

[0107] d(m 0 )=Ru(m 0 )=Ru 0 (m 0 )+Rδu(m 0 )

[0108] Among them, m 0is the initial velocity model, d is the full-wavefield data of the seismic data detected by the geophone, R represents the sampling operator of the geophone point, and Ru 0 (m 0 ) represents the transmitted wave in the seismic data, and Rδu(m 0 ) represents the primary reflected wave in the seismic data.

[0109] In one embodiment, in the target functional construction module, in the least squares sense, the target functional is constructed by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data. The target functional formula C(m 0 ) is as follows:

[0110]

[0111] Δd t = Ru 0 - d t

[0112] Δd r = Rδu - d r

[0113] where, Δd t is the data residual of the transmitted wave, Δd r is the data residual of the reflected wave, Ru 0 is the predicted transmitted wave data, d t is the observed transmitted wave data, Rδu is the predicted reflected wave data, and d r is the observed reflected wave data, and γ is the damping coefficient.

[0114] In one embodiment, the iterative update module is used to construct the inversion parameter gradient according to the target functional and using the model parameter gradient calculation formula. The model parameter gradient calculation formula is as follows:

[0115]

[0116] where, is the adjoint wavefield, is the adjoint perturbation wavefield.

[0117] In one embodiment, the wave equation waveform inversion processing device further includes a velocity field output module, which is used to output the low-to-medium wavenumber velocity field according to the iteratively updated initial velocity model.

[0118] The wave equation waveform inversion processing device comprises the initial velocity model establishment module, the corresponding relationship determination and splitting module, the target functional construction module, the inversion parameter gradient construction module, and the iterative update module, and adopts the wave equation waveform inversion processing method, including establishing an initial velocity model; determining the corresponding relationship between the full wave field data, the transmission wave data and the reflection wave data according to the initial velocity model, and splitting the original seismic shot gather record of the detected full wave field data into predicted transmission wave data and predicted reflection wave data according to the corresponding relationship; in the sense of least squares, the difference between the predicted reflection wave data and the observed reflection wave data and the difference between the predicted transmission wave data and the observed transmission wave data are minimized to construct the target functional; according to the target functional, the inversion parameter gradient is constructed by using the model parameter gradient calculation formula; the initial velocity model is iteratively updated with the obtained inversion parameter gradient to obtain the iteratively updated velocity model; in this way, based on the wave equation waveform inversion of the combination of the transmission wave and the reflection wave, the reflection wave has a deep lighting effect, which can be used to restore the medium and low wave number components in the deep part of the model, and can overcome the problem of insufficient long offset data to update the background model in the deep part, thereby improving the inversion accuracy. The above-mentioned wave equation waveform inversion processing method is based on the use of reflection wave information to update the deep part of the model. This technology combines the advantages of FWI and RWI (Reflection Wavelet Interferometry, reflection wave waveform inversion) inversion, establishes a joint inversion objective function of transmission wave and reflection wave to obtain accurate low-frequency background, and uses the low- and medium-wavenumber components of the shallow layer restored by the transmission wave waveform inversion, and uses the reflection wave waveform inversion to supplement the medium-wavenumber components of the medium- and deep-layer model. This application is particularly suitable for deep-water deep-ground velocity modeling problems. Starting from the transmission wave and reflection wave inversion, a transmission wave and reflection wave field joint waveform inversion based on high-precision wave theory is formed, forming a set of stable transmission wave and reflection wave joint full waveform inversion technology, and realizing high-precision velocity modeling of the shallow, medium and deep layers of the model. According to the difference in the wave paths of the transmission wave and the reflection wave, the present invention combines the corresponding path update direction, effectively makes up for the shortcomings of the traditional full waveform inversion, and realizes the simultaneous balanced modeling of the shallow, medium and deep layers of the model, which improves the inversion accuracy compared with the traditional method.

[0119] Example 4

[0120] In a third aspect, the present application provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the wave equation waveform inversion processing method as described in any of the above embodiments are implemented.

[0121] In one embodiment, a computer device is provided. The computer device may be a terminal, that is, a device to be upgraded. The internal structure diagram thereof may be as follows: Figure 5As shown. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used to communicate with the server through a network connection. When the computer program is executed by the processor, it realizes a method for wave equation waveform inversion processing. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, a touchpad, or a mouse, etc.

[0122] Those skilled in the art can understand that Figure 5 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0123] In one embodiment, when the processor executes the computer program, it also implements the following steps S110 to S150 of the wave equation waveform inversion processing method:

[0124] S110: Establish an initial velocity model;

[0125] In this embodiment, the initial velocity model can be established based on existing technical means. It should be noted that in the wave equation, the initial velocity model refers to the velocity distribution in the wave field at the initial moment. The change of the initial velocity model will affect the propagation and evolution of seismic waves. Therefore, for seismological research, it is very important to understand the initial velocity model. By analyzing and studying the initial velocity model, the structure and properties of the underground medium can be better understood, providing a reliable basis for research in fields such as earthquake prediction, mineral resource exploration, and environmental geology. It should be noted that the initial velocity model is only an approximate solution because in actual seismic observations, the propagation and evolution of seismic waves are affected by many factors, such as the inhomogeneity of the underground medium, complex terrain and landforms, and the characteristics of seismic sources. Therefore, it is necessary to conduct comprehensive analysis and interpretation in combination with other geological data and technical means.

[0126] In this embodiment, through the initial velocity model, the initial velocity model is iteratively updated and optimized based on transmitted waves and reflected waves later.

[0127] S120: Determine the corresponding relationships among the full-wavefield data, transmitted-wave data, and reflected-wave data according to the initial velocity model, and split the original seismic shot gather records of the detected full-wavefield data into predicted transmitted-wave data and predicted reflected-wave data according to the corresponding relationships;

[0128] In this embodiment, according to the established initial velocity model, determine the corresponding relationships among the full-wavefield data, transmitted-wave data, and reflected-wave data. In one embodiment, the determining the corresponding relationships among the full-wavefield data, transmitted-wave data, and reflected-wave data according to the initial velocity model is determined by using the following formula:

[0129] d(m 0 ) = Ru(m 0 ) = Ru 0 (m 0 ) + Rδu(m 0 )

[0130] where m 0 is the initial velocity model, d is the full-wavefield data of the seismic data detected by the geophone, R represents the sampling operator of the geophone point, Ru 0 (m 0 ) represents the transmitted wave in the seismic data, and Rδu(m 0 ) represents the primary reflected wave in the seismic data. In this way, by using the above formula, the corresponding relationships among the full-wavefield data, transmitted-wave data, and reflected-wave data can be determined better according to the initial velocity model.

[0131] In this example, considering that the underground disturbance has a certain continuity and spatial distribution, the superposition of the scattered wave fields forms the reflected wave (the scattered wave field mentioned below is the reflected wave). In seismic exploration, the seismic wave field propagates to the surface and is collected by a limited series of geophones to form the observed data (ignoring multiple scattering). For example, the reflected wave refers to the part of the wave that is reflected back into the original medium and continues to propagate when the wave encounters the medium interface. The transmitted wave is the part of the wave that passes through the interface and enters another medium and continues to propagate.

[0132] S130: In the sense of least squares, construct an objective functional by minimizing both the difference between the predicted reflected-wave data and the observed reflected-wave data and the difference between the predicted transmitted-wave data and the observed transmitted-wave data;

[0133] In this embodiment, the goal of seismic wave inversion is to find an appropriate solution in the model space to minimize the residual between the predicted data and the observed data. Based on the framework of the least squares sense, the objective functional is defined. In one embodiment, in the constructing the objective functional by minimizing both the difference between the predicted reflected-wave data and the observed reflected-wave data and the difference between the predicted transmitted-wave data and the observed transmitted-wave data in the sense of least squares, the objective functional formula C(m0 ) are as follows:

[0134]

[0135] Δd t = Ru 0 - d t

[0136] Δd r = Rδu - d r

[0137] where Δd t is the data residual of the transmitted wave, and Δd r is the data residual of the reflected wave, Ru 0 is the predicted transmitted wave data, d t is the observed transmitted wave data, Rδu is the predicted reflected wave data, and d r is the observed reflected wave data, and γ is the damping coefficient. Thus, by using the above objective functional formula, the differences between the predicted reflected wave data and the observed reflected wave data and between the predicted transmitted wave data and the observed transmitted wave data can be minimized in the least - squares sense.

[0138] S140: Construct the inversion parameter gradient according to the objective functional using the model parameter gradient calculation formula;

[0139] In this embodiment, according to the constructed objective functional, the inversion parameter gradient is constructed using the model parameter gradient calculation formula; the goal of the inversion is to find a suitable m 0 model such that the objective functional Cm 0 is minimized. In one embodiment, in the step of constructing the inversion parameter gradient according to the objective functional using the model parameter gradient calculation formula, the model parameter gradient calculation formula is as follows:

[0140]

[0141] where is the adjoint wavefield, is the adjoint perturbation wavefield. Comparing with the traditional FWI, the key of RWI is that it can better extract the reflection wave path information, and there is no high - frequency interference in the gradient formula. Thus, a joint inversion framework for the transmitted wave and the reflected wave of the wave equation can be better established.

[0142] S150: Iteratively update the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model.

[0143] In this embodiment, the obtained inversion parameter gradient is used to iteratively update the initial velocity model m 0Perform iterative updates to determine whether the velocity model after iterative updates meets the convergence condition; if it meets the convergence condition, obtain the velocity model after iterative updates, and if it does not meet the convergence condition, continue to repeat the foregoing steps. In this embodiment, the velocity model is updated based on the inversion framework, thereby ensuring that the mid- to low-wave number velocity field can be further determined subsequently.

[0144] In one of the embodiments, after the obtained inversion parameter gradient is used to perform iterative updates on the initial velocity model to obtain the velocity model after iterative updates, the wave equation waveform inversion processing method further includes the following steps:

[0145] Output the mid- to low-wave number velocity field according to the initial velocity model after iterative updates.

[0146] In this way, the mid- to low-wave number components in the deep part of the restoration model can be updated according to the initial velocity model, and the mid-wave number components of the mid- to deep-layer model can be supplemented by using the reflection wave waveform inversion, which can further improve the inversion accuracy.

[0147] The above-mentioned computer device, the processor executes the above-mentioned wave equation waveform inversion processing method, including establishing an initial velocity model; determining the corresponding relationship between the full wave field data, the transmission wave data and the reflection wave data according to the initial velocity model, and splitting the original seismic shot gather record of the detected full wave field data into predicted transmission wave data and predicted reflection wave data according to the corresponding relationship; in the sense of least squares, the difference between the predicted reflection wave data and the observed reflection wave data and the difference between the predicted transmission wave data and the observed transmission wave data are minimized to construct a target functional; according to the target functional, the inversion parameter gradient is constructed using the model parameter gradient calculation formula; the initial velocity model is iteratively updated with the obtained inversion parameter gradient to obtain an iteratively updated velocity model; in this way, based on the wave equation waveform inversion of the combination of the transmission wave and the reflection wave, the reflection wave has a deep lighting effect, which can be used to restore the medium and low wave number components in the deep part of the model, and can overcome the problem of insufficient long offset data to update the deep background model, thereby improving the inversion accuracy. The above-mentioned wave equation waveform inversion processing method is based on the use of reflection wave information to update the deep part of the model. This technology combines the advantages of FWI and RWI (Reflection Wavelet Interferometry, reflection wave waveform inversion) inversion, establishes the joint inversion objective function of transmission wave and reflection wave to obtain accurate low-frequency background, uses the low- and medium-wavenumber components of the shallow layer restored by the transmission wave waveform inversion, and uses the reflection wave waveform inversion to supplement the medium-wavenumber components of the medium- and deep-layer model. This application is particularly suitable for deep-water deep-earth velocity modeling problems. Starting from the transmission wave and reflection wave inversion, a transmission wave and reflection wave field joint waveform inversion based on high-precision wave theory is formed, forming a set of stable transmission wave and reflection wave joint full waveform inversion technology, and realizing high-precision velocity modeling of the shallow, medium and deep layers of the model. According to the difference in the wave paths of the transmission wave and the reflection wave, the present invention combines the corresponding path update direction, effectively makes up for the shortcomings of the traditional full waveform inversion, realizes the simultaneous balanced modeling of the shallow, medium and deep layers of the model, and improves the inversion accuracy compared with the traditional one.

[0148] In one embodiment, when the processor executes the computer program, it also implements the wave equation waveform inversion processing method and further performs the following steps:

[0149] After iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model, the wave equation waveform inversion processing method further comprises the following steps:

[0150] According to the iteratively updated initial velocity model, the medium and low wavenumber velocity fields are output.

[0151] In this way, the medium and low wavenumber components of the deep part of the model can be updated and restored according to the initial velocity model, and the medium wavenumber components of the deep part of the model can be supplemented by using the reflected wave waveform inversion, which can further improve the inversion accuracy.

[0152] The wave equation waveform inversion processing method comprises establishing an initial velocity model; determining the correspondence between full wave field data, transmission wave data and reflection wave data according to the initial velocity model, and splitting the original seismic shot gather record of the detected full wave field data into predicted transmission wave data and predicted reflection wave data according to the correspondence; constructing a target functional by minimizing the difference between the predicted reflection wave data and the observed reflection wave data and the difference between the predicted transmission wave data and the observed transmission wave data in the least square sense; constructing an inversion parameter gradient according to the target functional using a model parameter gradient calculation formula; iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain an iteratively updated velocity model; thus, based on the wave equation waveform inversion combined with the transmission wave and the reflection wave, the reflection wave has a deep illumination effect, can be used to restore the medium and low wave number components in the deep part of the model, can overcome the problem of insufficient long offset data to update the deep background model, and improve the inversion accuracy. The above-mentioned wave equation waveform inversion processing method is based on the use of reflection wave information to update the deep part of the model. This technology combines the advantages of FWI and RWI (Reflection Wavelet Interferometry, reflection wave waveform inversion) inversion, establishes a joint inversion objective function of transmission wave and reflection wave to obtain accurate low-frequency background, and uses the low- and medium-wavenumber components of the shallow layer restored by the transmission wave waveform inversion, and uses the reflection wave waveform inversion to supplement the medium-wavenumber components of the medium- and deep-layer model. This application is particularly suitable for deep-water deep-ground velocity modeling problems. Starting from the transmission wave and reflection wave inversion, a transmission wave and reflection wave field joint waveform inversion based on high-precision wave theory is formed, forming a set of stable transmission wave and reflection wave joint full waveform inversion technology, and realizing high-precision velocity modeling of the shallow, medium and deep layers of the model. According to the difference in the wave paths of the transmission wave and the reflection wave, the present invention combines the corresponding path update direction, effectively makes up for the shortcomings of the traditional full waveform inversion, and realizes the simultaneous balanced modeling of the shallow, medium and deep layers of the model, which improves the inversion accuracy compared with the traditional method.

[0153] Example 5

[0154] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the wave equation waveform inversion processing method as described in any of the above embodiments are implemented.

[0155] In one embodiment, when the processor executes the computer program, it also implements the following steps S110 to S150 of the wave equation waveform inversion processing method:

[0156] S110: Establishing an initial velocity model;

[0157] In this embodiment, an initial velocity model can be established based on existing technical means. It should be noted that in the wave equation, the initial velocity model refers to the velocity distribution in the wave field at the initial moment. The change of the initial velocity model will affect the propagation and evolution of seismic waves. Therefore, for seismic research, it is very important to understand the initial velocity model. By analyzing and studying the initial velocity model, we can better understand the structure and properties of the underground medium, providing a reliable basis for research in fields such as earthquake prediction, mineral resource exploration, and environmental geology. It should be noted that the initial velocity model is only an approximate solution because in actual seismic observations, the propagation and evolution of seismic waves are affected by many factors, such as the inhomogeneity of the underground medium, complex topography, and the characteristics of the seismic source. Therefore, it is necessary to conduct comprehensive analysis and interpretation in combination with other geological data and technical means.

[0158] In this embodiment, through the initial velocity model, the initial velocity model is iteratively updated and optimized based on transmitted waves and reflected waves later.

[0159] S120: According to the initial velocity model, determine the corresponding relationships among the full wavefield data, transmitted wave data, and reflected wave data, and split the original seismic shot gather record of the detected full wavefield data into predicted transmitted wave data and predicted reflected wave data according to the corresponding relationships;

[0160] In this embodiment, according to the established initial velocity model, determine the corresponding relationships among the full wavefield data, transmitted wave data, and reflected wave data. In one embodiment, the method for determining the corresponding relationships among the full wavefield data, transmitted wave data, and reflected wave data according to the initial velocity model is determined by the following formula:

[0161] d(m 0 ) = Ru(m 0 ) = Ru 0 (m 0 ) + Rδu(m 0 )

[0162] where m 0 is the initial velocity model, d is the full wavefield data of the seismic data detected by the geophone, R represents the sampling operator of the geophone point, Ru 0 (m 0 ) represents the transmitted wave in the seismic data, and Rδu(m 0 ) represents the primary reflected wave in the seismic data. Thus, by using the above formula, the corresponding relationships among the full wavefield data, transmitted wave data, and reflected wave data can be better determined based on the initial velocity model.

[0163] In this example, considering that the underground disturbance has a certain continuity and spatial distribution, the superposition of the scattered wave fields forms the reflected waves (the scattered wave fields mentioned below are the reflected waves). In seismic exploration, the seismic wave field propagates to the surface and is collected by a limited series of geophones to form the observed data (multiple scattering is ignored). For example, a reflected wave refers to the part of the wave that is reflected back into the original medium and continues to propagate when the wave encounters the interface of the medium. A transmitted wave is the part of the wave that passes through the interface and enters another medium and continues to propagate when the wave passes through the interface of the medium.

[0164] S130: In the least squares sense, construct the objective functional by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data;

[0165] In this embodiment, the goal of seismic wave inversion is to find an appropriate solution in the model space to minimize the residual between the predicted data and the observed data. Based on the framework of the least squares sense, the objective functional is defined. In one embodiment, in the step of constructing the objective functional by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data in the least squares sense, the objective functional formula C(m 0 ) is as follows:

[0166]

[0167] Δd t = Ru 0 - d t

[0168] Δd r = Rδu - d r

[0169] where Δd t is the data residual of the transmitted wave, Δd r is the data residual of the reflected wave, Ru 0 is the predicted transmitted wave data, d t is the observed transmitted wave data, Rδu is the predicted reflected wave data, d r is the observed reflected wave data, and γ is the damping coefficient. Thus, by using the above objective functional formula, it is possible to minimize both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data in the least squares sense.

[0170] S140: According to the objective functional, construct the inversion parameter gradient using the model parameter gradient calculation formula;

[0171] In this embodiment, according to the constructed objective functional, the inversion parameter gradient is constructed by using the model parameter gradient calculation formula; the goal of inversion is to find a suitable m 0 model such that the objective functional Cm 0 is minimized. In one embodiment, in the step of constructing the inversion parameter gradient according to the objective functional by using the model parameter gradient calculation formula, the model parameter gradient calculation formula is as follows:

[0172]

[0173] wherein, is the adjoint wave field, is the adjoint perturbation wave field. Comparing with the traditional FWI, it can be seen that the key of RWI is that it can better extract the reflection wave path information, and there is no high-frequency interference in the gradient formula. In this way, a joint inversion framework of the transmission wave and the reflection wave of the wave equation can be established better.

[0174] S150: Iteratively update the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model.

[0175] In this embodiment, the obtained inversion parameter gradient is used to iteratively update the initial velocity model m 0 to determine whether the iteratively updated velocity model meets the convergence condition; if it meets the convergence condition, the iteratively updated velocity model is obtained, and if it does not meet the convergence condition, the foregoing steps are repeated. In this embodiment, the velocity model is updated based on the inversion framework, so as to further ensure that the medium and low wavenumber velocity field can be determined subsequently.

[0176] In one embodiment, after the step of iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model, the wave equation waveform inversion processing method further includes the following steps:

[0177] Output the medium and low wavenumber velocity field according to the iteratively updated initial velocity model.

[0178] In this way, the medium and low wavenumber components in the deep part of the model can be updated and restored according to the initial velocity model, and the medium wavenumber components of the medium and deep layer models can be inversely supplemented by using the reflection wave waveform, which can further improve the inversion accuracy.

[0179] The above-mentioned computer device, the processor executes the above-mentioned wave equation waveform inversion processing method, including establishing an initial velocity model; determining the corresponding relationship between the full wave field data, the transmission wave data and the reflection wave data according to the initial velocity model, and splitting the original seismic shot gather record of the detected full wave field data into predicted transmission wave data and predicted reflection wave data according to the corresponding relationship; in the sense of least squares, the difference between the predicted reflection wave data and the observed reflection wave data and the difference between the predicted transmission wave data and the observed transmission wave data are minimized to construct a target functional; according to the target functional, the inversion parameter gradient is constructed using the model parameter gradient calculation formula; the initial velocity model is iteratively updated with the obtained inversion parameter gradient to obtain an iteratively updated velocity model; in this way, based on the wave equation waveform inversion of the combination of the transmission wave and the reflection wave, the reflection wave has a deep lighting effect, which can be used to restore the medium and low wave number components in the deep part of the model, and can overcome the problem of insufficient long offset data to update the deep background model, thereby improving the inversion accuracy. The above-mentioned wave equation waveform inversion processing method is based on the use of reflection wave information to update the deep part of the model. This technology combines the advantages of FWI and RWI (Reflection Wavelet Interferometry, reflection wave waveform inversion) inversion, establishes the joint inversion objective function of transmission wave and reflection wave to obtain accurate low-frequency background, uses the low- and medium-wavenumber components of the shallow layer restored by the transmission wave waveform inversion, and uses the reflection wave waveform inversion to supplement the medium-wavenumber components of the medium- and deep-layer model. This application is particularly suitable for deep-water deep-earth velocity modeling problems. Starting from the transmission wave and reflection wave inversion, a transmission wave and reflection wave field joint waveform inversion based on high-precision wave theory is formed, forming a set of stable transmission wave and reflection wave joint full waveform inversion technology, and realizing high-precision velocity modeling of the shallow, medium and deep layers of the model. According to the difference in the wave paths of the transmission wave and the reflection wave, the present invention combines the corresponding path update direction, effectively makes up for the shortcomings of the traditional full waveform inversion, realizes the simultaneous balanced modeling of the shallow, medium and deep layers of the model, and improves the inversion accuracy compared with the traditional one.

[0180] In one embodiment, when the processor executes the computer program, it also implements the wave equation waveform inversion processing method and further performs the following steps:

[0181] After iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain the iteratively updated velocity model, the wave equation waveform inversion processing method further comprises the following steps:

[0182] According to the iteratively updated initial velocity model, the medium and low wavenumber velocity fields are output.

[0183] In this way, the medium and low wavenumber components of the deep part of the model can be updated and restored according to the initial velocity model, and the medium wavenumber components of the deep part of the model can be supplemented by using the reflected wave waveform inversion, which can further improve the inversion accuracy.

[0184] The processor of the computer-readable storage medium executes the wave equation waveform inversion processing method, including establishing an initial velocity model; determining the correspondence between full wave field data, transmission wave data and reflection wave data according to the initial velocity model, and splitting the original seismic shot gather record of the detected full wave field data into predicted transmission wave data and predicted reflection wave data according to the correspondence; in the sense of least squares, the difference between the predicted reflection wave data and the observed reflection wave data and the difference between the predicted transmission wave data and the observed transmission wave data are minimized to construct a target functional; according to the target functional, the inversion parameter gradient is constructed using the model parameter gradient calculation formula; the initial velocity model is iteratively updated with the obtained inversion parameter gradient to obtain an iteratively updated velocity model; in this way, based on the wave equation waveform inversion of the combination of transmission wave and reflection wave, the reflection wave has a deep lighting effect, which can be used to restore the medium and low wave number components in the deep part of the model, and can overcome the problem of insufficient long offset data to update the deep background model, thereby improving the inversion accuracy. The above-mentioned wave equation waveform inversion processing method is based on the use of reflection wave information to update the deep part of the model. This technology combines the advantages of FWI and RWI (Reflection Wavelet Interferometry, reflection wave waveform inversion) inversion, establishes the joint inversion objective function of transmission wave and reflection wave to obtain accurate low-frequency background, uses the low- and medium-wavenumber components of the shallow layer restored by the transmission wave waveform inversion, and uses the reflection wave waveform inversion to supplement the medium-wavenumber components of the medium- and deep-layer model. This application is particularly suitable for deep-water deep-earth velocity modeling problems. Starting from the transmission wave and reflection wave inversion, a transmission wave and reflection wave field joint waveform inversion based on high-precision wave theory is formed, forming a set of stable transmission wave and reflection wave joint full waveform inversion technology, and realizing high-precision velocity modeling of the shallow, medium and deep layers of the model. According to the difference in the wave paths of the transmission wave and the reflection wave, the present invention combines the corresponding path update direction, effectively makes up for the shortcomings of the traditional full waveform inversion, realizes the simultaneous balanced modeling of the shallow, medium and deep layers of the model, and improves the inversion accuracy compared with the traditional one.

[0185] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0186] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification. It should be noted that the phrases "in one embodiment of the present application", "for example", "again, for example", etc. are intended to illustrate the present application rather than limit it. The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent application. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A method for wave equation waveform inversion processing, characterized in that, it includes the following steps: Establish an initial velocity model; According to the initial velocity model, determine the corresponding relationships of the full wavefield data, transmitted wave data, and reflected wave data, and split the original seismic shot gather record of the detected full wavefield data into predicted transmitted wave data and predicted reflected wave data according to the corresponding relationships; Under the least squares sense, construct an objective functional by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data; According to the objective functional, construct the inversion parameter gradient using the model parameter gradient calculation formula; Iteratively update the initial velocity model with the obtained inversion parameter gradient to obtain an iteratively updated velocity model.

2. The wave equation waveform inversion processing method according to claim 1, characterized in that, the step of determining the corresponding relationships of the full wavefield data, transmitted wave data, and reflected wave data according to the initial velocity model is determined by the following formula: d(m 0 ) = Ru(m 0 ) = Ru 0 (m 0 ) + Rδu(m 0 ) where m 0 is the initial velocity model, d is the full-wavefield data of the seismic data detected by the geophone, R represents the sampling operator of the geophone point, and Ru 0 (m 0 ) represents the transmitted wave in the seismic data, and Rδu(m 0 ) represents the primary reflected wave in the seismic data.

3. The wave equation waveform inversion processing method according to claim 2, characterized in that, In the least squares sense, the objective functional is constructed by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data. The objective functional formula C(m 0 ) is as follows: Δd t = Ru 0 - d t . Δd r = Rδu - d r . Among them, Ad t is the data residual of the transmitted wave, and Δd r is the data residual of the reflected wave. Ru 0 is the predicted transmitted wave data, d t is the observed transmitted wave data, Rδu is the predicted reflected wave data, and d r is the observed reflected wave data, and γ is the damping coefficient.

4. The wave equation waveform inversion processing method according to claim 3, characterized in that, in the step of constructing the inversion parameter gradient using the model parameter gradient calculation formula according to the objective functional, the model parameter gradient calculation formula is as follows: Among them, is the adjoint wave field, is the adjoint perturbation wave field.

5. The wave equation waveform inversion processing method according to any one of claims 1-4, characterized in that, after the step of iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain an iteratively updated velocity model, the wave equation waveform inversion processing method further includes the following steps: Output a low and medium wavenumber velocity field according to the iteratively updated initial velocity model.

6. A wave equation waveform inversion processing device, characterized in that, the device includes: An initial velocity model establishment module for establishing an initial velocity model; A corresponding relationship determination and splitting module for determining the corresponding relationships of the full wavefield data, transmitted wave data, and reflected wave data according to the initial velocity model, and splitting the original seismic shot gather record of the detected full wavefield data into predicted transmitted wave data and predicted reflected wave data according to the corresponding relationships; An objective functional construction module for constructing an objective functional by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data under the least squares sense; An inversion parameter gradient construction module for constructing the inversion parameter gradient using the model parameter gradient calculation formula according to the objective functional; An iterative update module for iteratively updating the initial velocity model with the obtained inversion parameter gradient to obtain an iteratively updated velocity model.

7. The wave equation waveform inversion processing device according to claim 6, characterized in that, in the corresponding relationship determination and splitting module, the step of determining the corresponding relationships of the full wavefield data, transmitted wave data, and reflected wave data according to the initial velocity model is determined by the following formula: d(m 0 ) = Ru(m 0 ) = Ru 0 (m 0 ) + Rδu(m 0 ) where m 0 is the initial velocity model, d is the full-wavefield data of the seismic data detected by the geophone, R represents the sampling operator of the geophone point, and Ru 0 (m 0 ) represents the transmitted wave in the seismic data, and Rδu(m 0 ) represents the primary reflected wave in the seismic data.

8. The wave equation waveform inversion processing device according to claim 7, characterized in that, In the target functional construction module, in the least squares sense, the target functional is constructed by minimizing both the difference between the predicted reflected wave data and the observed reflected wave data and the difference between the predicted transmitted wave data and the observed transmitted wave data. The target functional formula C(m 0 ) is as follows: Δd t = Ru 0 - d t Δd r = Rδu - d r where, Δd t is the data residual of the transmitted wave, Δd r is the data residual of the reflected wave, Ru 0 is the predicted transmitted wave data, d t is the observed transmitted wave data, Rδu is the predicted reflected wave data, d r is the observed reflected wave data, and γ is the damping coefficient.

9. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, when the processor executes the program, the steps of the wave equation waveform inversion processing method according to any one of claims 1-5 are implemented.

10. A computer-readable storage medium, on which a computer program is stored, wherein, when the computer program is executed by a processor, the steps of the wave equation waveform inversion processing method according to any one of claims 1-5 are implemented.

Images