A pre-stack AVA elastic parameter inversion method and system based on knowledge transfer

Abstract

The present invention discloses a pre-stack AVA elastic parameter inversion method and system based on knowledge transfer. A forward model is constructed based on the exact Zoeppritz equation to implement multiple groups of differential evolution algorithms with mutation. By introducing the optimal solution information of the elastic parameters of the bypass channel through the idea of knowledge transfer, while accelerating the iteration convergence of the current channel, the lateral continuity of the inverted elastic parameter profile is ensured. By introducing the post-stack P-wave impedance information during the selection process of multiple groups of differential evolution algorithms with mutation, the relationship between parameters is constrained, and the correlation of the three inverted parameters is improved; thereby ultimately improving the accuracy of pre-stack AVA elastic parameter inversion and ensuring lateral continuity. The basis of the present invention is for the known pre-stack AVA forward angle gather in the model, and the single-channel inversion of elastic parameters is performed using the MMDE method based on the given initial model. Then, to ensure lateral continuity, knowledge transfer is used to map the results of the bypass channel to the inversion process of the current channel, and finally, a laterally continuous elastic parameter inversion result is obtained.

Description

Technical Field

[0001] The present invention belongs to the technical field of geophysical exploration, and particularly relates to a pre-stack AVA elastic parameter inversion method and system based on knowledge transfer. Background Art

[0002] Energy is the most fundamental driving force for the development of the entire world and economic growth, and is the basis for human survival. Conventional energy sources such as oil and natural gas are the main energy sources for promoting social progress and civilization development. With the rapid development of society, the demand for oil and gas resources is also increasing. Among them, seismic exploration is an effective method for oil exploration. It can utilize the differences in the elasticity and density of underground media, and adopt signal processing and inversion methods to observe and analyze the propagation law of seismic waves generated by artificial earthquakes underground, predict reservoir parameters, thereby infer the properties and geometric structures of underground rock formations, and be used to predict underground oil and gas enrichment areas. Therefore, the research and development of various seismic exploration technologies are the main research issues in the field of geophysical exploration.

[0003] As one of the main methods of seismic pre-stack inversion, AVA (Amplitude Variation with Angle) inversion uses the relationship between the reflection amplitude and the incident angle to search for oil and gas resources, and can simultaneously obtain reservoir parameters such as P-wave velocity, S-wave velocity, and density. It has become one of the hot technologies in the research of oil and gas reservoir prediction, and has played an important role in the fields of oil and gas detection and reservoir characterization. It directly analyzes seismic pre-stack data, makes full use of the original seismic information obtained by multiple coverage, can obtain more elastic parameters, and is more conducive to the identification of reservoir lithology and fluid properties. At the same time, using the formation elastic parameters extracted from the actual seismic trace data in the pre-stack data can also help predict lithology and hydrocarbons and conduct quantitative reservoir description. Pre-stack elastic parameter inversion has strong ill-posedness and there are multi-parameter coupling problems. The traditional pre-stack AVA parameter inversion method uses the generalized linear inversion method to transform the non-linear least squares problem into a least squares problem of parameter correction amounts, and simplifies the non-linear algorithm of the Zoeppritz equation, which ensures the inversion accuracy to a certain extent. However, it also has problems such as falling into local extrema, difficult parameter selection, and difficult convergence.

[0004] To improve the accuracy of the inverted elastic parameters and reduce the dependence on the initial model, the AVA parameter inversion method based on intelligent optimization algorithms has received attention, and some work has been proposed. For example, the particle swarm optimization algorithm based on quantum behavior directly applies the global optimization algorithm to the AVA elastic parameter inversion. In the process of elastic parameter inversion based on the genetic algorithm and ray theory, empirical relations are used as constraint conditions to analyze the correlation between the longitudinal wave, transverse wave velocity, and density in the recovered parameters. However, the above methods do not consider the lateral continuity of elastic parameters, and there are also problems such as lack of accuracy in the constraint relationship of empirical formulas. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a pre-stack AVA elastic parameter inversion method and system based on knowledge transfer for solving the technical problem of poor lateral continuity in the inversion profile results in traditional pre-stack AVA elastic parameter inversion.

[0006] The present invention adopts the following technical solutions:

[0007] A pre-stack AVA elastic parameter inversion method based on knowledge transfer, comprising the following steps:

[0008] S1. Partially intercept the post-stack Marmousi II impedance model, generate a shear wave velocity and density model profile using the shale line empirical formula, perform forward modeling on the P-wave and S-wave velocity and density model data using the exact Zoeppritz equation, and convolve with a Ricker wavelet with a dominant frequency of 30 Hz to obtain pre-stack seismic data;

[0009] S2. Perform low-frequency filtering on the P-wave and S-wave velocity and density model data obtained in step S1 according to the multi-group differential evolution algorithm with mutation to generate an initial population;

[0010] S3. At a specific number of iterations G, apply knowledge transfer to the optimal solution P of the adjacent trace obtained by iterative inversion of the logging trace or adjacent seismic trace and the optimal solution Q of the current trace at the current number of iterations G to provide candidate solutions that ensure lateral continuity for population mutation Use the candidate solutions obtained by knowledge transfer and the individuals in the population of the current trace at the G-th iteration to perform mutation and crossover to obtain a population incorporating the optimal solution information of the adjacent trace, accelerating the iterative convergence process of the current trace;

[0011] S4. During the selection process of the multi-group differential evolution algorithm with mutation, according to the set threshold, use the post-stack impedance information to constrain the correlation between the P-wave velocity and density, perform EPS filtering on the candidate solutions in the initial population obtained in step S2, and compare the objective function value J[V p ,Vs , ρ], to obtain the individual with the minimum objective function value at the current iteration;

[0012] S5. Take the individual with the minimum objective function value at the current iteration in step S4 as the optimal solution Q G Guide the evolution of the mapping matrix M in knowledge transfer. Set the maximum number of iterations, and iterate steps S2 - S4. When the maximum number of iterations is reached, take the individual in the population with the minimum objective function value as the optimal parameter solution, and at the same time as the bypass optimal solution for the next knowledge transfer, that is, obtain the optimal elastic parameter inversion result V for the current channel p , V s , ρ; at the same time, take the optimal solution obtained in the current channel as the bypass optimal solution P used in the knowledge transfer module during the inversion process of the next channel.

[0013] Specifically, in step S1, generate shear wave and density profiles using the empirical formula as follows:

[0014]

[0015] where, v s,i is the S-wave velocity value of the i-th channel, rho i is the density value of the i-th channel, v p,i is the density value of the i-th channel.

[0016] Specifically, in step S1, the seismic data d is:

[0017]

[0018] where, is convolution, w is the wavelet moment used in forward modeling, and r is the P-P wave reflection coefficient obtained according to the Zoeppritz equation.

[0019] Specifically, in step S2, the filter used for low-frequency filtering is a Butterworth low-frequency filter, with a cut-off frequency of 3 Hz and the order set to 2. In single-channel inversion, the initial population is randomly generated within ±20% of the low-frequency elastic parameters after filtering.

[0020] Specifically, in step S3, the candidate solution obtained by knowledge transfer is:

[0021]

[0022] where, P is the bypass optimal solution vector, M G is the mapping matrix M at the G-th iteration, ΔG is the iteration interval, is the mapping operator, and M is the mapping matrix.

[0023] Furthermore, the solution process of the mapping matrix M is as follows:

[0024]

[0025] Among them, Q G is the optimal solution of the current inversion trace at the G-th iteration, and P is the optimal solution vector of the adjacent trace.

[0026] Furthermore, during the single-trace inversion process, after multiple tests, the iteration interval ΔG is set to 20 in the model experiment.

[0027] Specifically, in step S4, the objective function constrained by the post-stack impedance information is:

[0028]

[0029] Among them, threshold is a preset threshold, is the post-stack P-wave impedance information of the current trace, is the P-wave velocity information of the current trace during the iterative inversion process, and ρ cur is the density information of the current trace during the iterative inversion process.

[0030] Specifically, in step S5, the objective function J[V p ,V s ,ρ] is:

[0031]

[0032] Among them, v p ,v s ,ρ are elastic parameters, G(·) is a forward operator, and d obs is the observed data.

[0033] In the second aspect, an embodiment of the present invention provides a pre-stack AVA elastic parameter inversion system based on knowledge transfer, including:

[0034] A forward modeling module that partially intercepts the post-stack Marmousi II impedance model, generates a shear wave velocity and density model profile using the shale line empirical formula, performs forward modeling on the P-wave and S-wave velocity and density model data using the exact Zoeppritz equation, and convolves with a Ricker wavelet with a dominant frequency of 30 Hz to obtain pre-stack seismic data;

[0035] A filtering module that performs low-frequency filtering on the P-wave and S-wave velocity and density model data obtained by the forward modeling module according to a multi-group differential evolution algorithm with mutation to generate an initial population;

[0036] A migration module that, at a specific iteration number G, applies knowledge transfer to the optimal solution P of the adjacent trace obtained by iterative inversion of the logging trace or adjacent seismic trace and the optimal solution Q of the current trace at the current iteration number GAbove, provide candidate solutions that ensure horizontal continuity for population variation Candidate solutions obtained by knowledge transfer Mutate and crossover with the individuals in the population of the current trace at the G-th iteration to obtain a population incorporating the optimal solution information of the adjacent traces, accelerating the iterative convergence process of the current trace;

[0037] Function module. During the selection process of the multi-group mutation differential evolution algorithm, according to the set threshold, use the post-stack impedance information to constrain the correlation between the P-wave velocity and density, perform EPS filtering on the candidate solutions in the initial population obtained by the migration module, and compare the objective function values J[V p ,V s ,ρ], and obtain the individual with the minimum objective function value at the current iteration;

[0038] Inversion module. Take the individual with the minimum objective function value at the current iteration of the function module as the optimal solution Q G Guide the evolution of the mapping matrix M in knowledge transfer, set the maximum number of iterations, perform iterations, and when the maximum number of iterations is reached, take the individual in the population with the minimum objective function value as the parameter optimal solution, and at the same time as the optimal solution of the adjacent trace for the next trace's knowledge transfer, that is, obtain the optimal elastic parameter inversion result V p ,V s ,ρ of the current trace; At the same time, take the optimal solution obtained by the current trace as the optimal solution P of the adjacent trace used in the knowledge transfer module during the inversion process of the next trace.

[0039] Compared with the prior art, the present invention has at least the following beneficial effects:

[0040] A pre-stack AVA elastic parameter inversion method based on knowledge transfer constructs a forward model based on the exact Zoeppritz equation, implements a multi-group mutation differential evolution algorithm based on MATLAB, and in the mutation stage of the multi-group mutation differential evolution algorithm, incorporates the optimal solution information of the adjacent trace into the candidate solutions of the population of the current trace through the mapping operator of knowledge transfer, and incorporates the post-stack P-wave impedance information into the selection stage of the multi-group mutation differential evolution algorithm. Through the constraint of the post-stack P-wave impedance information, the nonlinear relationship between the P-wave velocity and density is constrained, further accelerating the convergence rate of the global optimization process, reducing the nonlinearity of the inversion problem and improving the accuracy of the inversion of the three elastic parameters, so as to finally obtain a laterally continuous and relatively accurate elastic parameter profile result.

[0041] Furthermore, using the empirical formula to generate the S-wave velocity and density from the P-wave velocity is a necessary step to obtain the reflection coefficient using the exact Zoeppritz equation and forward model the pre-stack seismic data.

[0042] Furthermore, the seismic data d generated by forward modeling of P-wave and S-wave velocities and density is used as the known pre-stack seismic observation data for minimizing the inversion objective function through iterative inversion.

[0043] Furthermore, a Butterworth low-pass filter is used to perform low-pass filtering on the P-wave and S-wave velocities and density, aiming to obtain the low-frequency components used to initialize the population during the inversion process. Since it is necessary to ensure the diversity of individuals within the population during population initialization, each individual in the population is randomly initialized by setting the upper and lower bounds to ±20% of the elastic parameter low-frequency components.

[0044] Furthermore, traditional pre-stack AVA elastic parameter inversion methods all use single-trace inversion methods and do not consider the lateral continuity of elastic parameters of adjacent seismic traces. Therefore, in order to introduce the elastic parameter information of adjacent traces into the inversion process of the current seismic trace, candidate solutions are obtained through the knowledge transfer term. These candidate solutions are used as individuals in the population iterative inversion process of the current trace and participate in the mutation and crossover processes.

[0045] Furthermore, it is unreasonable to directly transfer the optimal solution information P of adjacent traces to the population of the current seismic trace to obtain candidate solutions by introducing the elastic parameter information of adjacent traces into the iterative inversion process of the current trace through knowledge transfer. Therefore, a mapping operator is required. The mapping matrix M is calculated, and the candidate solution individuals are obtained by processing the adjacent trace solution P through the mapping matrix M.

[0046] Furthermore, the reason for setting the iterative update interval ΔG of the mapping matrix M to 20 is that the number of iterative inversions of single-trace elastic parameters is usually large. If the mapping matrix M is updated at each iteration, it will bring a large amount of additional time overhead. And it has been proven by experiments that by setting the iterative update interval ΔG to 20, the additional calculation time overhead and the use of the optimal solution information P of adjacent traces can be well balanced.

[0047] Furthermore, the main purpose of introducing the post-stack impedance information constraint is to reduce the ill-posedness of pre-stack AVA elastic parameter inversion. Since pre-stack seismic data is jointly constrained by P-wave and S-wave velocities and density. Therefore, only using as the final inversion objective function will lead to non-uniqueness of the inverted elastic parameter results. Therefore, by introducing the post-stack impedance information constraint and using the physical relationship between P-wave velocity and density, the ill-posedness of the inversion problem can be well reduced.

[0048] Furthermore, the set J[V p ,V s ,ρ] is the objective function that needs to be minimized during the single-trace inversion process, and it is also the evaluation index for iteratively updating individuals in the population through the selection step during the iteration process. Through multiple iterations, J[V p ,Vs , ρ] The individual that obtains the minimum value is considered to be the elastic parameter V obtained by inversion p , V s , ρ.

[0049] It can be understood that for the beneficial effects of the second aspect above, reference can be made to the relevant descriptions in the first aspect above, and details will not be repeated here.

[0050] In summary, within the framework of the global optimization method, the present invention realizes a pre-stack AVA elastic parameter inversion method. By introducing the knowledge transfer term, this method overcomes the shortcomings of the traditional AVA elastic parameters, such as the inability to effectively utilize the side-channel seismic information and the lateral discontinuity of the inversion parameter profile. At the same time, by guiding the iterative inversion process of the current seismic trace with the elastic parameter information of the side-channel, the convergence is accelerated. In addition, for the inherent ill-posedness of the pre-stack AVA elastic parameter inversion, the ill-posedness of this inverse problem is reduced by the post-stack impedance information constraint, and the accuracy of the elastic parameter inversion result is improved.

[0051] Next, through the attached drawings and embodiments, the technical solution of the present invention will be further described in detail. Brief Description of the Drawings

[0052] Figure 1 It is a schematic flow chart of the present invention;

[0053] Figure 2 It is a single-channel AVA angle gather obtained by forward modeling with the Zoeppritz equation;

[0054] Figure 3 It is a single-channel inversion flow chart of the multi-group differential evolution algorithm with mutation;

[0055] Figure 4 It is a schematic diagram of the initial model of the three elastic parameters for inversion. Among them, (a) is the low-frequency longitudinal wave velocity profile, (b) is the low-frequency shear wave velocity profile, and (c) is the low-frequency density profile;

[0056] Figure 5 It is a schematic diagram of obtaining the candidate solution of this trace by knowledge transfer mapping;

[0057] Figure 6 It is a schematic diagram of the three elastic parameters obtained by the empirical formula. Among them, (a) is the true longitudinal wave velocity profile, (b) is the true shear wave velocity profile, and (c) is the true density profile;

[0058] Figure 7 It is a comparison chart of the single-channel convergence curves with and without knowledge transfer;

[0059] Figure 8Schematic diagrams of the angle gather and convergence curve of the single - trace inversion result. Among them, (a) is the Observed AVAData, (b) is the Inversion AVA Data, and (c) is the convergence curve of the 100th trace;

[0060] Figure 9 Comparison diagram of the single - trace elastic three - parameter inversion results with and without knowledge transfer. Among them, (a) is the P - Wave Velocity, (b) is the S - Wave Velocity, (c) is the Density, (d) is the P - Wave Velocity, (e) is the S - Wave Velocity, and (f) is the Density;

[0061] Figure 10 Schematic diagram of the elastic three - parameter profile inversion result with knowledge transfer. Among them, (a) is the inverted P - wave velocity profile, (b) is the inverted S - wave velocity profile, and (c) is the inverted density profile. Specific implementation manner

[0062] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0063] In the description of the present invention, it should be understood that the terms "include" and "comprise" indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.

[0064] It should also be understood that the terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an", and "the" are intended to include the plural forms.

[0065] It should be further understood that the term " / and / or" used in the specification of the present invention and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the preceding and following related objects.

[0066] It should be understood that although terms such as first, second, and third may be used in the embodiments of the present invention to describe preset ranges and the like, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from each other. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0067] Depending on the context, the word "if" as used herein can be interpreted as "when" or "while" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "if determined" or "if detected (stated condition or event)" can be interpreted as "when determined" or "in response to determining" or "when detected (stated condition or event)" or "in response to detecting (stated condition or event)".

[0068] Various structural schematic diagrams according to the disclosed embodiments of the present invention are shown in the drawings. These figures are not drawn to scale, where for the purpose of clear expression, some details are enlarged and some details may be omitted. The shapes of various regions and layers shown in the figures, as well as their relative sizes and positional relationships, are only exemplary, and in practice, there may be deviations due to manufacturing tolerances or technical limitations, and those skilled in the art can design regions / layers with different shapes, sizes, and relative positions according to actual needs.

[0069] The elastic properties in the subsurface medium, such as the P-wave velocity, S-wave velocity, and density, play an important role in subsequent processing such as target location of the subsurface medium, formation structure interpretation, and reservoir prediction, and ultimately affect the interpretability and reliability of the final results. As one of the main methods of seismic prestack inversion, AVA (Amplitude Variation with Angle) inversion uses the relationship between the reflection amplitude and the incident angle to search for oil and gas resources, and can obtain reservoir parameters such as the P-wave velocity, S-wave velocity, and density at the same time.

[0070] The present invention provides a pre-stack AVA elastic parameter inversion method based on knowledge transfer. A forward model is constructed based on the exact Zoeppritz equation, and a multi-group mutation differential evolution algorithm (MMDE) is implemented. In the multi-group mutation differential evolution algorithm (MMDE), the optimal solution information of the flank elastic parameters is introduced through the idea of knowledge transfer, which accelerates the iterative convergence of the current trace while ensuring the lateral continuity of the inverted elastic parameter profile. In addition, by introducing the post-stack P-wave impedance information during the selection process of the multi-group mutation differential evolution algorithm (MMDE), the relationship between parameters is constrained, and the correlation of the inverted three-parameter results is improved. Thus, the accuracy of the pre-stack AVA elastic parameter inversion is ultimately improved, and the lateral continuity is ensured. The basis of the present invention is the single-trace inversion of elastic parameters using the MMDE method for the known pre-stack AVA forward angle gathers in the model based on a given initial model. Then, to ensure lateral continuity, knowledge transfer is used to map the results of the flank traces to the inversion process of the current trace, and finally, a laterally continuous inverted elastic parameter result is obtained.

[0071] Please refer to Figure 1 , a pre-stack AVA elastic parameter inversion method based on knowledge transfer according to the present invention, comprising the following steps:

[0072] S1. Intercept the post-stack impedance model from the Marmousi II model, generate the P-wave and S-wave velocity and density models using empirical formulas, and perform forward modeling on the P-wave and S-wave velocity and density model data using the exact Zoeppritz equation to obtain seismic data;

[0073] The post-stack Marmousi II model has 100 points at a sampling interval of 2 ms in the y direction and 200 points in the x direction.

[0074] The S-wave and density profiles are generated using the empirical formulas as follows:

[0075]

[0076] where, v s,i is the S-wave velocity value of the i-th trace, and rho i is the density value of the i-th trace.

[0077] The magnitudes of the elastic parameters obtained from the empirical formulas are all 100×200, where 100 corresponds to the number of layers in the y direction and 200 corresponds to the number of seismic traces.

[0078] The exact Zoeppritz equation is:

[0079]

[0080] where, θ i (i = 1, 2, 3, 4) respectively represent the P-wave incident angle, P-wave transmission angle, SV-wave reflection angle, and transmission angle, and Rpp , R ps is the P-wave / S-wave reflection coefficient, T pp , T ps is the P-wave / S-wave transmission coefficient. Here, only the most commonly used P-wave reflection coefficient R is considered pp , α 1,2 , β 1,2 , ρ 1，2 are the P-wave and S-wave velocities and densities of the upper and lower media, respectively

[0081] According to Snell's theorem, it satisfies the following relationship:

[0082]

[0083] where p represents the ray parameter

[0084] Using the above exact Zoeppritz equation, we can obtain the P-wave reflection coefficient R pp , which, after convolution with the seismic wavelet, gives the forward seismic data d:

[0085]

[0086] where the symbol represents convolution, and during the forward process, with Δθ = 4°, 10 angles are uniformly selected from 0° to 40° as the forward angle values. The single-channel angle gather obtained by forward modeling is as Figure 2 shown, with a size of 100×10

[0087] S2. Implement the multi-group mutation differential evolution algorithm (MMDE) based on the MATLAB software platform. According to the multi-group mutation differential evolution algorithm, perform low-frequency filtering on the seismic data obtained in step S1 to generate the initial population;

[0088] Please refer to Figure 3 , the multi-group mutation differential evolution algorithm (MMDE) combines the cooperative multi-group mutation differential evolution algorithm (CCDE) and the differential evolution algorithm (DE) during the mutation process, and adds EPS filtering in the selection stage of the algorithm to ensure the block structure of elastic three-parameter inversion. The upper and lower bounds for generating the initial population in single-channel inversion are ±20% of the low frequency, as Figure 4 shown

[0089] The filter uses a Butterworth low-frequency filter with a cut-off frequency set to 3 Hz and an order set to 2. Perform low-frequency filtering on the three-parameter logging data using a Butterworth low-frequency filter, and use the filtered low-frequency information as the search space for the population

[0090] S3. At a specific number of iterations G, apply knowledge transfer to the flank optimal solution P and the optimal solution Q of the current channel at the current number of iterationsG Above, provide candidate solutions that ensure horizontal continuity for population variation Candidate solutions obtained by knowledge transfer And perform mutation and crossover processes on individuals in the population;

[0091] The knowledge transfer (Knowledge Transfer) method guides the transfer direction through a mapping matrix, that is:

[0092]

[0093] Among them, Q G Is the optimal solution of the current inversion trace at the G-th iteration, P is the vector of optimal solutions of adjacent traces, M is the mapping matrix, which is calculated by the mapping operator Calculated.

[0094] Determine when to multiply the mapping matrix M with the optimal solution P of the adjacent trace by a given iteration interval ΔG, as Figure 5 Shown.

[0095] The current trace solution vector finally introducing knowledge transfer is expressed as:

[0096]

[0097] Among them, Is the candidate solution individual after migration of the current inversion trace in the G-th generation.

[0098] During the inversion process of a single trace, the iteration interval ΔG is selected as 20 to avoid frequent modification of the mapping matrix M resulting in a decrease in inversion speed and negative transfer.

[0099] Candidate solutions obtained by knowledge transfer And the mutation and crossover processes of individuals in the population are specifically as follows:

[0100] Use the candidate solution obtained by migration at G = 20 As a solution in the current trace population to guide the mutation direction, and further modify the mapping matrix M with the obtained Q G To obtain an improved migration solution vector

[0101] S4. In the selection process of the multi-group mutation differential evolution algorithm (MMDE), use the post-stack impedance information to constrain the correlation between the P-wave velocity and density according to the set threshold (threshold), perform EPS filtering on the candidate solutions in the initial population obtained in step S2, and compare the objective function values;

[0102] The objective function constrained by the post-stack impedance information is:

[0103]

[0104] Among them, threshold is a preset threshold value, representing the post-stack P-wave impedance information of the current trace.

[0105] S5. Take the individual with the minimum fitness function at the current iteration number obtained in step S4 as the optimal solution Q G Guide the evolution of the mapping matrix M in knowledge transfer, avoid negative transfer, and iterate steps S2 - S4 according to the set maximum iteration number, return the individual with the minimum objective function value as the optimal solution of the parameters, and at the same time as the optimal solution P of the adjacent trace for the next knowledge transfer.

[0106] The objective function of the multi-group mutation differential evolution algorithm (MMDE) is:

[0107]

[0108] Among them, v p , v s , ρ are elastic parameters, G(·) is a forward modeling operator, which is implemented based on the exact Zoeppritz equation and the convolution model, d obs is the observed data.

[0109] By minimizing the L2 norm of the observed data and the forward modeled data, the optimal solutions of the three parameters of P-wave, S-wave velocity and density are obtained, and this process is repeated in the elastic parameters of each seismic trace, and finally the inversion result of the laterally continuous elastic parameter profile is obtained.

[0110] In another embodiment of the present invention, a pre-stack AVA elastic parameter inversion system based on knowledge transfer is provided. This system can be used to implement the above-mentioned pre-stack AVA elastic parameter inversion method based on knowledge transfer. Specifically, the pre-stack AVA elastic parameter inversion system based on knowledge transfer includes a forward modeling module, a filtering module, a transfer module, a function module and an inversion module.

[0111] Among them, the forward modeling module partially intercepts the post-stack Marmousi II impedance model, generates the S-wave velocity and density model profiles using the shale line empirical formula, performs forward modeling on the P-wave and S-wave velocity and density model data using the exact Zoeppritz equation, and convolves with a Ricker wavelet with a dominant frequency of 30 Hz to obtain pre-stack seismic data;

[0112] The filtering module performs low-frequency filtering on the P-wave and S-wave velocity and density model data obtained by the forward modeling module according to the multi-group mutation differential evolution algorithm to generate an initial population;

[0113] The transfer module, at a specific iteration number G, applies knowledge transfer to the optimal solution P of the adjacent trace obtained by iterative inversion of the logging trace or adjacent seismic traces and the optimal solution Q of the current iteration number of the current traceG provide candidate solutions that ensure horizontal continuity for population variation Candidate solutions obtained by knowledge transfer Mutate and crossover with the individuals in the population of the current trace at the G-th iteration to obtain a population incorporating the optimal solution information of the adjacent traces, accelerating the iterative convergence process of the current trace;

[0114] Function module. During the selection process of the multi-group mutation differential evolution algorithm, according to the set threshold, use the post-stack impedance information to constrain the correlation between the P-wave velocity and density, perform EPS filtering on the candidate solutions in the initial population obtained by the transfer module, and compare the objective function values J[V p ,V s ,ρ], to obtain the individual with the minimum objective function value at the current iteration;

[0115] Inversion module, take the individual with the minimum objective function value at the current iteration of the function module as the optimal solution Q G Guide the evolution of the mapping matrix M in knowledge transfer, set the maximum number of iterations, and perform iterations. When the maximum number of iterations is reached, take the population individual with the minimum objective function value as the parameter optimal solution, and at the same time as the adjacent trace optimal solution for the next trace knowledge transfer, that is, obtain the optimal elastic parameter inversion result V p ,V s ,ρ; At the same time, take the optimal solution obtained in the current trace as the adjacent trace optimal solution P used in the knowledge transfer module during the inversion process of the next trace.

[0116] In another embodiment of the present invention, a terminal device is provided. The terminal device includes a processor and a memory. The memory is used to store a computer program. The computer program includes program instructions. The processor is used to execute the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to implement the corresponding method process or corresponding function; The processor described in the embodiment of the present invention can be used for the operation of the pre-stack AVA elastic parameter inversion method based on knowledge transfer, including:

[0117] Partially intercept the stacked Marmousi II impedance model, and use the empirical formula of shale line to generate the shear wave velocity and density model profiles. Use the exact Zoeppritz equation to forward model the P-wave and S-wave velocity and density model data, and convolve with a Ricker wavelet with a dominant frequency of 30 Hz to obtain pre-stack seismic data; perform low-frequency filtering on the obtained P-wave and S-wave velocity and density model data according to multiple sets of differential evolution algorithms with mutation to generate an initial population; at a specific number of iterations G, apply knowledge transfer to the optimal solution P of the adjacent trace obtained by iterative inversion of the logging trace or adjacent seismic trace and the optimal solution Q of the current trace at the current number of iterations G to provide candidate solutions that ensure lateral continuity for population mutation Candidate solutions obtained by using knowledge transfer and the individuals in the population at the G-th iteration of the current trace are mutated and crossed to obtain a population incorporating the optimal solution information of the adjacent trace, accelerating the iterative convergence process of the current trace; during the selection process of multiple sets of differential evolution algorithms with mutation, according to the set threshold, use the post-stack impedance information to constrain the correlation between the P-wave velocity and density, and perform EPS filtering on the candidate solutions in the obtained initial population, and compare the objective function values J[V p ,V s ,ρ] to obtain the individual with the minimum objective function value at the current number of iterations; take the individual with the minimum objective function value at the current number of iterations as the optimal solution Q G Guide the evolution of the mapping matrix M in knowledge transfer, set the maximum number of iterations, and perform iterations. When the maximum number of iterations is reached, take the individual of the population with the minimum objective function value as the parameter optimal solution, and at the same time as the optimal solution of the adjacent trace for knowledge transfer of the next trace, that is, obtain the optimal elastic parameter inversion result V p ,V s ,ρ; at the same time, take the optimal solution obtained by the current trace as the optimal solution P of the adjacent trace used in the knowledge transfer module during the inversion process of the next trace.

[0118] In another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory). The computer-readable storage medium is a memory device in a terminal device and is used to store programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and, of course, the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, and the operating system of the terminal is stored in this storage space. And, one or more instructions suitable for being loaded and executed by the processor are also stored in this storage space. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory (Non-Volatile Memory), such as at least one disk memory.

[0119] One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the pre-stack AVA elastic parameter inversion method based on knowledge transfer in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor to perform the following steps:

[0120] Partially intercept the post-stack Marmousi II impedance model, and use the shale line empirical formula to generate the shear wave velocity and density model profiles. Use the exact Zoeppritz equation to perform forward modeling on the P-wave and S-wave velocity and density model data, and convolve with a Ricker wavelet with a main frequency of 30 Hz to obtain pre-stack seismic data; perform low-frequency filtering on the obtained P-wave and S-wave velocity and density model data according to multiple sets of differential evolution algorithms with mutations to generate an initial population; at a specific number of iterations G, apply knowledge transfer to the optimal solution P of the side channel obtained by iterative inversion of the logging trace or adjacent seismic traces and the optimal solution Q of the current trace at the current number of iterations G to provide candidate solutions that ensure lateral continuity for population mutation Candidate solutions obtained by using knowledge transfer and individuals in the population of the current trace at the G-th iteration are mutated and crossed to obtain a population incorporating the optimal solution information of the side channel, accelerating the iterative convergence process of the current trace; during the selection process of multiple sets of differential evolution algorithms with mutations, according to the set threshold, use the post-stack impedance information to constrain the correlation between the P-wave velocity and density, perform EPS filtering on the candidate solutions in the obtained initial population, and compare the objective function values J[V p ,V s ,ρ], and obtain the individual with the minimum objective function value at the current iteration; use the individual with the minimum objective function value at the current iteration as the optimal solution Q GGuide the evolution of the mapping matrix M in knowledge transfer, set the maximum number of iterations, and perform iterations. When the maximum number of iterations is reached, take the population individual with the minimum objective function value as the optimal parameter solution, and at the same time as the bypass optimal solution for the next knowledge transfer, that is, obtain the optimal elastic parameter inversion result V of the current channel p , V s , ρ; at the same time, take the optimal solution obtained in the current channel as the bypass optimal solution P adopted in the knowledge transfer module during the inversion process of the next channel.

[0121] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the present invention described and shown in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0122] Implementation example

[0123] The specific implementation process of the present invention is applied to the Marmousi II model. This model has 100 points at a sampling interval of 2 ms in the y direction and 200 points in the x direction. The elastic three-parameter model for method testing is as follows Figure 6 shown. Among them Figure 6 (a), Figure 6 (b), and Figure 6 (c) correspond to the P-wave velocity model, S-wave velocity model, and density model respectively.

[0124] The initial model obtained by filtering with a Butterworth low-pass filter with a cut-off frequency of 3 Hz is as follows Figure 4 shown. Among them Figure 4 (a), Figure 4 (b), and Figure 4 (c) correspond to the P-wave velocity low-frequency model, S-wave velocity low-frequency model, and density low-frequency model respectively. They correspond to Figure 6 (a), Figure 6 (b), and Figure 6 (c) respectively. Figure 2 Shown in is the 100th angle gather after forward modeling using the Zoeppritz equation. It has 100 points in the y direction with a sampling interval of 2 ms and 10 angles in the x direction, corresponding to 4° to 40° respectively.

[0125] The elastic three-parameter inversion is performed on the 200 pre-stack AVA angle gathers obtained from the forward modeling, and the obtained angle gathers and convergence curves are as follows Figure 8 shown. Among them Figure 8 (a) shows the AVA angle gather obtained from the forward modeling using the true elastic three parameters at the 100th trace, while Figure 8 (b) shows the AVA angle gather obtained from the forward modeling of the elastic three parameters after the algorithm iterates 200 times. It can be seen that the algorithm has achieved good results in data matching. Figure 8 (c) is the convergence curve during the algorithm iteration process. It can be seen that during the 200 iterations, the objective function value of the algorithm has decreased by one order of magnitude and finally converges to about 0.04, which indicates the good performance of the algorithm and verifies its applicability to solving the strongly non-linear problem of elastic three parameters.

[0126] To further illustrate the improvement in lateral continuity brought by knowledge transfer, Figure 7 and Figure 9 respectively compare the convergence curves and single-trace elastic three-parameter results with and without adding the knowledge transfer term during the section inversion process. Among them, Figure 7 compares the convergence curves with and without adding the knowledge transfer term in the inversion of the same trace. It can be seen that after adding the knowledge transfer term, the descent speed of the convergence curve is significantly faster than that without adding the knowledge transfer term, which indicates the positive effect of the knowledge transfer term on the rapid convergence of this trace. Figure 9 (a), Figure 9 (b), Figure 9 (c) are the inversion results of the 100th trace with the knowledge transfer term added, while Figure 9 (d), Figure 9 (e), Figure 9 (f) are the single-trace inversion results without adding the knowledge transfer term. Among them Figure 9 (a) and Figure 9 (d) comparison can clearly show that there is an obvious value mismatch in the inversion result of the longitudinal wave velocity at the low-velocity anomaly at 0.12 s without adding the knowledge transfer term. Figure 9 (b) and Figure 9 (e) comparison of the inversion results of the shear wave velocity also shows a similar phenomenon. Figure 9 (c) and Figure 9 (f) have a greater difference in the inversion results of the density term because the influence weight of the density term on the AVA angle gather is smaller and its non-linearity is stronger. After adding the knowledge transfer term, the density term can also be well inverted. This also shows that the knowledge transfer term can play an implicit regularization effect.

[0127] Figure 10 (a), Figure 10 (b), Figure 10(c) shows the profile inversion results of the elastic three parameters. From top to bottom, they are the profile inversion results of the P-wave velocity, the profile inversion results of the S-wave velocity, and the profile inversion results of the density. Comparing it with Figure 6 the true elastic three-parameter profile, it can be found that the algorithm well restores the structure of the elastic three parameters, and it can also well invert the structure of the low-velocity anomaly body between the 80th and 120th traces. And due to the introduction of knowledge transfer and post-stack impedance constraint, the inversion results of the elastic parameters are guaranteed in both lateral continuity and accuracy, which further proves the effectiveness of the pre-stack AVA elastic parameter intelligent inversion method based on knowledge transfer in solving nonlinear problems and its high accuracy in the three-parameter inversion.

[0128] In summary, the present invention provides a pre-stack AVA elastic parameter inversion method and system based on knowledge transfer, which can consider the influence of the side-channel information on the inversion result of the current trace during the inversion process of the elastic parameter profile. Therefore, compared with the single-trace inversion, it can better ensure the lateral continuity, and the side-channel information can accelerate the convergence of the current trace, providing the effect of implicit regularization. The post-stack P-wave impedance constraint further constrains the relationship between the elastic three parameters. Therefore, the proposed method has the ability to improve the inversion accuracy of the elastic three parameters and ensure the lateral continuity.

[0129] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the above-mentioned division of each functional unit and module is used as an example. In actual applications, the above-mentioned functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of the present application. The specific working processes of the units and modules in the above system can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated here.

[0130] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0131] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0132] In the embodiments provided by the present invention, it should be understood that the disclosed device / terminal and method can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of the device or unit can be in an electrical, mechanical or other form.

[0133] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0134] In addition, the functional units in each embodiment of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0135] When the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above-described embodiment methods of the present invention, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-described method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0136] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device generate a device for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0137] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0138] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are executed on the computer or other programmable apparatus to produce a computer-implemented process, thereby the instructions executed on the computer or other programmable apparatus provide steps for realizing the functions specified in one process or a plurality of processes and / or blocks Figure 1 one process or a plurality of processes and / or blocks Figure 1 and / or steps for realizing the functions specified in one block or a plurality of blocks.

[0139] The above is only to illustrate the technical idea of the present invention, and the protection scope of the present invention cannot be limited thereby. Any modifications made on the basis of the technical solution according to the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A prestack AVA elastic parameter inversion method based on knowledge transfer, characterized in that It includes the following steps: S1. Partially intercept the stacked Marmousi II impedance model, generate the shear wave velocity and density model profiles using the shale line empirical formula, perform forward modeling on the P-wave and S-wave velocity and density model data using the exact Zoeppritz equation, and convolve with a Ricker wavelet with a dominant frequency of 30 Hz to obtain pre-stack seismic data; S2. Perform low-frequency filtering on the P-wave and S-wave velocity and density model data obtained in step S1 according to the multi-group differential evolution algorithm with mutation to generate the initial population; S3. At a specific number of iterations G, apply knowledge transfer to the optimal solution P of the side channel obtained by iterative inversion of the logging trace or adjacent seismic traces and the optimal solution Q of the current trace at the current number of iterations, G to provide candidate solutions that ensure lateral continuity for population mutation. Candidate solutions obtained by using knowledge transfer and individuals in the population of the current trace at the G-th iteration undergo mutation and crossover to obtain a population incorporating the optimal solution information of the side channel, accelerating the iterative convergence process of the current trace; S4. During the selection process of multiple groups of mutation differential evolution algorithms, according to the set threshold, using the post-stack impedance information to constrain the correlation between the P-wave velocity and density, perform EPS filtering on the candidate solutions in the initial population obtained in step S2, and compare the objective function values J[V p , V s , ρ], and obtain the individual with the minimum objective function value at the current iteration number; S5. Take the individual with the minimum objective function value at the current iteration of step S4 as the optimal solution Q G Guide the evolution of the mapping matrix M in knowledge transfer, set the maximum number of iterations, and iterate steps S2 - S4. When the maximum number of iterations is reached, take the individual in the population with the minimum objective function value as the optimal parameter solution and, at the same time, as the bypass optimal solution for the next knowledge transfer, that is, obtain the optimal elastic parameter inversion result V of the current channel p ,V s , ρ; at the same time, take the optimal solution obtained in the current channel as the bypass optimal solution P used in the knowledge transfer module during the inversion process of the next channel 2. The pre-stack AVA elastic parameter inversion method based on knowledge transfer according to claim 1, characterized in that In step S1, the shear wave and density profiles are generated using the empirical formula as follows: where, v s,i is the S-wave velocity value of the i-th trace, rho i is the density value of the i-th trace, v p,i is the density value of the i-th trace.

3. The pre-stack AVA elastic parameter inversion method based on knowledge transfer according to claim 1, characterized in that In step S1, the seismic data d is: wherein, is convolution, w is the wavelet moment used in forward modeling, and r is the P-P wave reflection coefficient obtained from the Zoeppritz equation.

4. The prestack AVA elastic parameter inversion method based on knowledge transfer according to claim 1, wherein, In step S2, the filter used for low-frequency filtering is a Butterworth low-pass filter with a cut-off frequency of 3 Hz and an order of 2. In the single-channel inversion, the initial population is randomly generated within ±20% of the low-frequency elastic parameters after filtering.

5. The pre-stack AVA elastic parameter inversion method based on knowledge transfer according to claim 1, wherein In step S3, the candidate solutions obtained by knowledge transfer are as follows: Among them, P is the optimal solution vector of the side channel, and M G is the mapping matrix M at the G-th iteration, ΔG is the iteration interval, is the mapping operator, and M is the mapping matrix.

6. The pre-stack AVA elastic parameter inversion method based on knowledge transfer according to claim 5, characterized in that The solution process of the mapping matrix M is as follows: Among them, Q G is the optimal solution of the current inversion channel at the G-th iteration, and P is the optimal solution vector of the side channel.

7. The pre-stack AVA elastic parameter inversion method based on knowledge transfer according to claim 5, characterized in that During the single-channel inversion process, after multiple tests, the iteration interval ΔG is set to 20 in the model experiment.

8. The pre-stack AVA elastic parameter inversion method based on knowledge transfer according to claim 1, characterized in that In step S4, the objective function constrained by the post-stack impedance information is: where threshold is a preset threshold value, is the post-stack P-wave impedance information of the current trace, is the P-wave velocity information of the current trace during the iterative inversion process, ρ cur is the density information of the current trace during the iterative inversion process.

9. The pre-stack AVA elastic parameter inversion method based on knowledge transfer according to claim 1, characterized in that In step S5, the objective function J[V p ,V s ,ρ] is as follows: where, v p , v s , ρ is the elastic parameter, G(·) is the forward operator, and d obs is the observed data.

10. A prestack AVA elastic parameter inversion system based on knowledge transfer, characterized in that It includes: A forward modeling module that partially intercepts the stacked Marmousi II impedance model, generates the shear wave velocity and density model profiles using the shale line empirical formula, performs forward modeling on the P-wave and S-wave velocity and density model data using the exact Zoeppritz equation, and convolves with a Ricker wavelet with a dominant frequency of 30 Hz to obtain pre-stack seismic data; A filtering module that performs low-frequency filtering on the P-wave and S-wave velocity and density model data obtained by the forward modeling module according to the multi-group differential evolution algorithm with mutation to generate the initial population; Migration module, at a specific number of iterations G, applies knowledge migration to the optimal solution P of the side channel obtained by iterative inversion of the logging trace or adjacent seismic traces and the optimal solution Q of the current trace at the current number of iterations, to provide candidate solutions that guarantee lateral continuity for population mutation. G Candidate solutions obtained by using knowledge migration and individuals in the population of the current trace at the G-th iteration undergo mutation and crossover to obtain a population incorporating the optimal solution information of the side channel, accelerating the iterative convergence process of the current trace; Function module, in the process of selecting the multi-group mutation differential evolution algorithm, according to the set threshold, using the post-stack impedance information to constrain the correlation between the P-wave velocity and density, performing EPS filtering on the candidate solutions in the initial population obtained by the migration module, and comparing the objective function values J[V p ,V s ,ρ], to obtain the individual with the minimum objective function value at the current iteration number; The inversion module takes the individual with the minimum objective function value at the current iteration of the function module as the optimal solution Q G Guide the evolution of the mapping matrix M in knowledge transfer, set the maximum number of iterations, and perform iterations. When the maximum number of iterations is reached, take the population individual with the minimum objective function value as the parameter optimal solution and also as the bypass optimal solution for the next knowledge transfer, that is, obtain the optimal elastic parameter inversion result V of the current channel p ,V s , ρ; at the same time, take the optimal solution obtained in the current channel as the bypass optimal solution P adopted in the knowledge transfer module during the inversion process of the next channel

Images