Seismic wave impedance inversion method and device, medium and equipment

Abstract

The invention provides a seismic wave impedance inversion method and device, a medium and equipment, and belongs to the technical field of seismic exploration. The method comprises the following steps: S1, establishing a seismic wave impedance inversion objective function; s2, extracting a reflection coefficient sequence by adopting a quantum particle swarm algorithm based on a cloud model; and S3, inverting the seismic trace wave impedance. Compared with a conventional wave impedance inversion method, the wave impedance inversion method provided by the invention is high in convergence speed, high in global optimization capability, high in inversion result precision and high in anti-noise capability, and is a feasible wave impedance inversion method.

Description

Technical Field

[0001] The present invention belongs to the technical field of seismic exploration, and in particular relates to a seismic wave impedance inversion method, device, medium and equipment. Background Art

[0002] Seismic wave impedance is a bridge between seismic and logging data. It plays a huge role in geophysical data interpretation and plays an important role in seismic exploration. Moreover, seismic wave impedance inversion belongs to solving optimization problems. Its optimization objective function may contain multiple continuous variables within a certain range. Traditional optimization methods have problems such as difficulty in optimizing some functions, easy to fall into local solutions and slow convergence speed. The nonlinear swarm intelligent optimization algorithm only considers the objective function, has low dependence on the initial model, and can randomly search in the global domain. It is one of the effective methods to solve wave impedance inversion problems.

[0003] The particle swarm algorithm is a random algorithm constructed by simulating the foraging behavior of biological groups. The quantum particle swarm algorithm is an improved algorithm proposed by combining it with quantum behavior characteristics. It has the advantages of simple evolution equations, few control parameters, and fast convergence speed. The cloud model can complete the uncertain conversion between qualitative concepts and quantitative descriptions. It is an innovation and development of the concept of membership function in fuzzy theory. Combining it with the quantum particle swarm, the algorithm has strong global search capabilities, high search accuracy and particle aggregation, and has been partially applied in the fields of intelligent control and data mining. However, this method is still rare in the field of seismic impedance inversion. Summary of the invention

[0004] The purpose of the present invention is to solve the problems existing in the above-mentioned prior art and to provide a seismic wave impedance inversion method, device, medium and equipment, which adopts a quantum particle swarm algorithm based on a cloud model to extract the reflection coefficient sequence, so that the wave impedance inversion method has the advantages of global optimization, high precision and fast algorithm.

[0005] The present invention is achieved through the following technical solutions:

[0006] A first aspect of the present invention provides a seismic impedance inversion method, which specifically comprises the following steps:

[0007] S1. Establishing the seismic wave impedance inversion objective function;

[0008] S2, using quantum particle swarm algorithm based on cloud model to extract reflection coefficient sequence;

[0009] S3. Invert seismic trace impedance.

[0010] A further improvement of the present invention is:

[0011] In step S1, the objective function is established using the least squares principle, which is expressed as follows:

[0012]

[0013] Where: N is the number of samples, S(t) is the seismic record sequence, W(t) is the seismic wavelet sequence, R(t) is the seismic reflection coefficient sequence, and * represents convolution.

[0014] A further improvement of the present invention is:

[0015] The specific operations in step S2 include:

[0016] Given a particle population size of N, the iteration process is the tth step, the particle moves in a D-dimensional space, and the potential well of the particle in the dth dimension is p id (t), the update equation of particle x(t) is described as:

[0017]

[0018]

[0019]

[0020] Among them, P i (t) is the current optimal position of the particle, G(t) and C(t) are the global optimal position and average optimal position of the population; and u(t) are uniformly distributed random numbers in the interval [0,1]; β is the contraction and expansion factor.

[0021] A further improvement of the present invention is:

[0022] Particle P i The update method of G(t) and G(t) adopts the following strategy:

[0023]

[0024]

[0025] A further improvement of the present invention is:

[0026] The cloud model is used to adapt to generate the shrinkage and expansion factors. The normal cloud model is represented by three values: expected value Ex, entropy En, and super entropy He.

[0027] make

[0028]

[0029] Then there is

[0030]

[0031] Among them, f g 、f w 、f i Respectively represent the optimal, worst fitness and individual fitness of particles in the population;

[0032] Because 0i It is closely related to the particle fitness, which is mapped to the desired ideal interval to construct the contraction and expansion factor β as follows:

[0033]

[0034] k is a control parameter.

[0035] A further improvement of the present invention is:

[0036] Logistic chaotic mapping is used to generate the original population of the particle swarm algorithm. The specific expression is:

[0037] x n+1 =rx n (1-x n ) (11)

[0038] Among them, r is the chaos initiator;

[0039] In formula (11), when r = 4, the system is in a completely chaotic state. n+1 ∈(0,1).

[0040] A further improvement of the present invention is:

[0041] In step S3, seismic trace impedance is inverted, and the specific operations include:

[0042] When the wave impedance of the first layer of stratum is known, the discrete formula of wave impedance is:

[0043]

[0044] Q i , Q i+1 Represent the wave impedance values ​​of the i-th layer and the i+1-th layer respectively;

[0045] Substituting the extracted reflection coefficient sequence into equation (12) can invert the wave impedance of the seismic trace.

[0046] A second aspect of the present invention provides a seismic wave impedance inversion device, comprising:

[0047] A construction unit for establishing a seismic wave impedance inversion objective function;

[0048] An extraction unit, used for extracting a reflection coefficient sequence by using a quantum particle swarm algorithm based on a cloud model;

[0049] Inversion unit, invert seismic trace impedance.

[0050] According to a third aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores at least one computer-executable program, and when the at least one program is executed by the computer, the computer executes the steps in the seismic wave impedance inversion method as described above.

[0051] A fourth aspect of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps in the seismic wave impedance inversion method as described above.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] Compared with conventional wave impedance inversion methods, the wave impedance inversion method proposed in the present invention has fast convergence speed, strong global optimization capability, high inversion result accuracy, and strong anti-noise capability, and is a practical wave impedance inversion method. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 The present invention provides a flow chart of a seismic wave impedance inversion method;

[0055] Figure 2 is the wave impedance model and inversion results;

[0056] Figure 3 It is the convergence curve of wave impedance inversion using quantum particle swarm algorithm based on cloud model;

[0057] Figure 4 is the P-wave velocity and density curve of the actual data;

[0058] Figure 5 is the seismic wavelet used in the inversion;

[0059] Figure 6 is the extracted reflection coefficient sequence;

[0060] Figure 7 It is the seismic record obtained by convolution of wavelet and seismic reflection coefficient;

[0061] Figure 8 It is a comparison chart between the inverted wave impedance and the real wave impedance. DETAILED DESCRIPTION

[0062] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0063] [Example 1]

[0064] The embodiment of the present invention provides a seismic wave impedance inversion method, such as Figure 1 As shown, the specific steps include:

[0065] S1. Establishing the seismic wave impedance inversion objective function;

[0066] In seismic exploration, the mathematical model of seismic response can be expressed as:

[0067] S=W*R+δ (1)

[0068] Where: S is the seismic record sequence, W is the seismic wavelet sequence, R is the seismic reflection coefficient sequence, * represents convolution, and δ is a balanced white noise sequence that obeys Gaussian distribution.

[0069] When establishing the objective function here, balanced white noise is not considered. The objective function is established using the least squares principle and is expressed as follows:

[0070]

[0071] Where: N is the number of samples, and other parameters are the same as formula (1). In the model, W(t) is the known sub-wave, and R(t) is the required reflection coefficient; or R(t) is the known reflection coefficient, and W(t) is the required sub-wave.

[0072] If the convolution of the reflection coefficient and wavelet (or wavelet and reflection coefficient) obtained by the inversion algorithm and the sum of squares of the actual seismic record residuals reaches the set objective function value, then this reflection coefficient (or wavelet) is the desired ideal reflection coefficient (or wavelet).

[0073] It can be seen that the problem of extracting reflection coefficients (or wavelets) is essentially an optimization problem, that is, to match the synthetic seismic records with the actual seismic records so that the record residual is minimized. Due to the use of the least squares principle, this method has a certain anti-interference ability.

[0074] S2, using the quantum particle swarm algorithm based on the cloud model to extract the reflection coefficient sequence R(t);

[0075] The particle swarm algorithm is a random algorithm constructed by simulating the foraging behavior of biological groups. The quantum particle swarm algorithm is an improved algorithm proposed by combining it with quantum behavior characteristics. It has the advantages of simple evolution equations, few control parameters, and fast convergence speed. The cloud model can complete the uncertain conversion between qualitative concepts and quantitative descriptions. It is an innovation and development of the concept of membership function in fuzzy theory. Combining it with the quantum particle swarm, the algorithm has a strong global search capability, high search accuracy and particle aggregation, and has been partially applied in the fields of intelligent control and data mining.

[0076] In quantum space, particles do not have a definite trajectory for movement, and the state of particles is specifically described by the wave function Ψ(x, t). First, the probability density function of the particle appearing at a certain point is obtained by solving the Schrödinger equation, and then the particle's position equation is obtained through Monte Carlo simulation.

[0077] Given a particle population size of N, the iteration process is the tth step, the particle moves in D-dimensional space, and the potential well of the particle in the dth dimension is p id (t), the update equation of particle x(t) can be described as:

[0078]

[0079]

[0080]

[0081] Among them, P i (t) is the current optimal position of the particle, G(t) and C(t) are the global optimal position and average optimal position of the population; and u(t) is a uniformly distributed random number in the interval [0,1]; β is the shrinkage and expansion factor, which is the only control parameter of this algorithm. Common selection methods are fixed value, linear or nonlinear value selection methods.

[0082] Particle P i The update method of G(t) and G(t) is similar to other intelligent algorithms, that is, the following strategy is adopted:

[0083]

[0084]

[0085] Preferably, a cloud model adaptation is used to generate the shrink-expansion factors.

[0086] The normal cloud model is a set of random numbers that follows the normal distribution law and has a stable tendency. It is represented by three values: expected value Ex, entropy En, and super entropy He. Their respective meanings are:

[0087] Expected value Ex: the point in the number space that best represents the qualitative concept, reflecting the center of gravity of the cloud;

[0088] Entropy En: On the one hand, it reflects the range of language values ​​that can be accepted in the number domain space, and on the other hand, it also reflects the probability that a point in the number domain space can represent this language value. It represents the randomness of the cloud droplets of qualitative concepts, and it reveals the correlation between fuzziness and randomness.

[0089] Hyperentropy He: The uncertainty measure of entropy, that is, the entropy of entropy, reflects the cohesion of uncertainty of all points representing the language value in the number domain space, that is, the cohesion of cloud droplets.

[0090] make

[0091]

[0092] Then there is

[0093]

[0094] Among them, f g 、f w 、f i They represent the optimal, worst fitness and individual fitness of particles in the population respectively.

[0095] Because 0i It is closely related to the particle fitness, which can be mapped to the desired ideal interval to construct the contraction-expansion factor β as follows:

[0096]

[0097] k is a control parameter, and its value also affects the final convergence and convergence accuracy of the algorithm. (10) In the formula, for individuals with different fitness, the β obtained in the next iteration is different. The individuals with good fitness will get a larger β, and vice versa, thus achieving the purpose of adaptive control.

[0098] For points close to the current global optimal position of the particle swarm, the value of β should be relatively large to make the particles diverge as much as possible, which is conducive to expanding the search range and jumping out of the local extreme point; for points far away from the global optimal position of the particle swarm, the value of β should be relatively small to make the particles converge, maintain a large range of optimization, and accelerate the convergence of the algorithm.

[0099] Furthermore, the value range of β can be adjusted by adjusting the control parameter k. For example, better optimization effect can be obtained by controlling it in the interval [0.6 0.8].

[0100] Prioritize, Logistic chaos mapping is used to generate the original population of the particle swarm algorithm. Logistic chaos mapping is a method commonly used to give initial values ​​in intelligent swarming algorithms. Compared with uniformly distributed random number generators, chaos mapping algorithms can make the initial random numbers more consistent with the characteristics of "randomness". Therefore, when the number of global iterations is not very large, this method has the most uniform numerical distribution, and its structure is simple and insensitive to initial values. The specific expression is:

[0101] x n+1 =rx n (1-x n ) (11)

[0102] Among them, r is the chaos initiator.

[0103] Furthermore, in formula (11), when r=4, the system is in a completely chaotic state. n+1 ∈(0,1).

[0104] In this algorithm, the input and output are both the positions of particles. The input is the initial position of the particle, and the output is the updated position calculated by the algorithm.

[0105] The initial position of a given particle is the initial reflection coefficient, and the position after the algorithm updates represents the reflection coefficient sequence after the algorithm solves it. In other words, the x in the algorithm becomes the reflection coefficient R when solving the reflection coefficient sequence problem.

[0106] S3, inversion of seismic trace wave impedance;

[0107] When the wave impedance of the first layer is known, the discrete formula for wave impedance (the product of density and velocity) is:

[0108]

[0109] Q i , Q i+1 Represent the impedance values ​​of the i-th layer and the i+1-th layer respectively;

[0110] Substituting the extracted reflection coefficient sequence into equation (12), the wave impedance of the seismic trace can be inverted.

[0111] [Example 2]

[0112] Given an 8-layer theoretical model reflection coefficient sequence, by convolution with the theoretical wavelet (using the Ricker wavelet, the main frequency is 35Hz), according to formula (1), the wave impedance values ​​(through) are [5,10,18,4,11,8,40,45] (unit: 10 9 g / s*m 2 ),like Figure 2 Indicated by the thicker dotted line.

[0113] The objective function of formula (2) is iterated. The number of particle swarms is 80, the particle space dimension is 20, the initial value of particle x (the initial position of the particle) is given by formula (11), the shrinkage and expansion factor β is given by formulas (8)-(10) and limited to the range of [0.6, 0.8], the number of iterations is set to 10000 times, and the convergence threshold is controlled to 1e -8 .

[0114] According to the steps introduced in the present invention, the wave impedance is inverted, and the convergence process is as follows: Figure 3The inversion results are shown in Figure 2 Indicated by the thin dotted line.

[0115] from Figure 2 Comparison of the inversion results with the real model shows that the seismic wave impedance obtained by inversion is very consistent with the given real seismic wave impedance. Whether in places with larger impedance values ​​or smaller impedance values, the impedance obtained by inversion is highly consistent with the given model impedance and does not weaken with increasing depth, proving that the method of the present invention is an effective wave impedance inversion method.

[0116] From the convergence curve, the overall trend is rapidly decreasing, indicating that this cloud model-based quantum particle swarm algorithm has the ability to search globally. In addition, the error was reduced to below 0.01 at the beginning of the iteration (less than 10 times), indicating that this method has the characteristics of fast optimization. As the number of iterations increases, the inversion accuracy is further improved, indicating that this method has the characteristics of high accuracy in addition to its fast speed.

[0117] [Example 3]

[0118] In order to better illustrate the application process of the present invention to actual data, Example 3 is given.

[0119] This embodiment is aimed at actual well logging data, and the P-wave velocity and density values ​​are read from the well logging data, such as Figure 4 shown.

[0120] Since it is difficult for actual data to have zero phase, the Ormsby wavelet that is more in line with the characteristics of actual data is used. Since the low-frequency information in the wavelet has a greater impact on the inversion result, its low-frequency information is retained here, and the Ormsby wavelet sequence containing low-frequency information is generated by computer simulation, such as Figure 5 shown.

[0121] For well logging data, since there is information about velocity, density and depth, it is equivalent to giving the information of each small stratum underground, and based on this, the reflection coefficient sequence is given, such as Figure 6 This is equivalent to the true reflection coefficient sequence.

[0122] Using formula (1), Figure 5 , Figure 6 The wavelet and reflection coefficient sequence can be used to obtain synthetic seismic records such as Figure 7 shown.

[0123] Using step 2 of the present invention, an initial reflection coefficient sequence is generated by randomly giving initial values, and the reflection coefficient sequence is updated by a quantum particle swarm method based on a cloud model. The updated reflection coefficient sequence and the given wavelet are used to obtain an inverted seismic record. The inverted seismic record is gradually approximated to the synthesized seismic record by the objective function shown in formula (2). The reflection coefficient sequence corresponding to each seismic record is recorded, and the wave impedance of the seismic trace is obtained by formula (12).

[0124] The specific parameters in the process of the present invention are similar to those in Example 2. Since the stratigraphic information of the actual data is richer, the number of iterations here is set to 100,000. The obtained wave impedance inversion results are compared with the wave impedance results of the actual logging, as shown in FIG. Figure 8 It can be seen that the wave impedance inverted by the method of the present invention is basically the same as the real wave impedance, which proves that the method has the characteristics of strong global convergence, fast speed, high precision, etc., and is an effective method for wave impedance inversion.

[0125] [Example 4]

[0126] The embodiment of the present invention provides a seismic wave impedance inversion device, comprising:

[0127] A construction unit for establishing a seismic wave impedance inversion objective function;

[0128] An extraction unit, used for extracting a reflection coefficient sequence by using a quantum particle swarm algorithm based on a cloud model;

[0129] Inversion unit, invert seismic trace impedance.

[0130] [Example 5]

[0131] This embodiment provides a computer-readable storage medium, which stores at least one computer-executable program. When the at least one program is executed by the computer, the computer executes the steps in the seismic wave impedance inversion method as described above.

[0132] [Example 6]

[0133] An embodiment of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps in the seismic wave impedance inversion method as described above.

[0134] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0135] The above technical solution is only one implementation mode of the present invention. For those skilled in the art, it is easy to make various types of improvements or modifications based on the principles disclosed in the present invention, and it is not limited to the technical solution described in the above specific embodiments of the present invention. Therefore, the above description is only preferred and does not have a restrictive meaning.

Claims

1. A seismic impedance inversion method, It is characterized in that The specific steps include: S1. Establishing the seismic wave impedance inversion objective function; S2, using quantum particle swarm algorithm based on cloud model to extract reflection coefficient sequence; S3. Invert seismic trace impedance.

2. The seismic wave impedance inversion method according to claim 1, It is characterized in that In step S1, the objective function is established using the least squares principle, which is expressed as follows: Where: N is the number of samples, S(t) is the seismic record sequence, W(t) is the seismic wavelet sequence, R(t) is the seismic reflection coefficient sequence, and * represents convolution.

3. The seismic wave impedance inversion method according to claim 1, It is characterized in that The specific operations in step S2 include: Given a particle population size of N, the iteration process is the tth step, the particle moves in a D-dimensional space, and the potential well of the particle in the dth dimension is p id (t), the update equation of particle x(t) is described as: Among them, P i (t) is the current optimal position of the particle, G(t) and C(t) are the global optimal position and average optimal position of the population; and u(t) are uniformly distributed random numbers in the interval [0,1]; β is the contraction and expansion factor.

4. The seismic wave impedance inversion method according to claim 3, It is characterized in that Particle P i The update method of G(t) and G(t) adopts the following strategy:

5. The seismic wave impedance inversion method according to claim 4, It is characterized in that The cloud model is used to adapt to generate the shrinkage and expansion factors. The normal cloud model is represented by three values: expected value Ex, entropy En, and super entropy He. make Ex=f g , Then there is Among them, f g 、f w 、f i Respectively represent the optimal, worst fitness and individual fitness of particles in the population; Because 0i It is closely related to the particle fitness, which is mapped to the desired ideal interval to construct the contraction and expansion factor β as follows: k is a control parameter.

6. The seismic wave impedance inversion method according to claim 5, It is characterized in that Logistic chaotic mapping is used to generate the original population of the particle swarm algorithm. The specific expression is: x n+1 =rx n (1-x n ) (11) Among them, r is the chaos initiator; In formula (11), when r = 4, the system is in a completely chaotic state. n+1 ∈(0,1).

7. The seismic wave impedance inversion method according to claim 1, It is characterized in that In step S3, the seismic wave impedance is inverted, and the specific operations include: When the wave impedance of the first layer of stratum is known, the discrete formula of wave impedance is: Q i , Q i+1 Represent the wave impedance values ​​of the i-th layer and the i+1-th layer respectively; Substituting the extracted reflection coefficient sequence into equation (12) can invert the wave impedance of the seismic trace.

8. A seismic wave impedance inversion device, It is characterized in that include: A construction unit for establishing a seismic wave impedance inversion objective function; An extraction unit, used for extracting a reflection coefficient sequence by using a quantum particle swarm algorithm based on a cloud model; Inversion unit, invert seismic trace impedance.

9. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores at least one computer-executable program, and when the at least one program is executed by the computer, the computer executes the steps in the seismic wave impedance inversion method according to any one of claims 1 to 7.

10. A computer device, It is characterized in that It comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps in the seismic wave impedance inversion method as described in any one of claims 1 to 7.

Images