Solid-liquid decoupling fluid elastic impedance inversion method considering extrusion spraying effect

Abstract

The invention relates to the technical field of rock physical modeling, in particular to a solid-liquid decoupling fluid elastic impedance inversion method considering a squeezing and spraying effect. Comprising the following steps: S1, constructing a solid-liquid decoupling fluid factor; s2, establishing a relational expression of a solid-liquid decoupling fluid factor and elastic impedance; s3, fluid factor inversion is carried out based on the Bayesian theory; on the basis of the elastic fluctuation theory of a Biot double-hole medium, the influence of local flow on the pore scale is considered, the extrusion jet flow effect between rock cracks is added, a new fluid indication factor Kfs is constructed, and a five-parameter elastic impedance inversion equation is deduced according to the new fluid indication factor Kfs. According to the solid-liquid decoupling fluid elastic impedance inversion method considering the extrusion spraying effect, the situation that Vp, Vs and rho are firstly obtained through conventional inversion, and then accumulated errors of fluid factors are calculated from the three parameters is avoided. Theoretical analysis and verification of actual data show that the fluid factor has high precision and sensitivity on detection of oil gas in a reservoir.

Description

Technical Field

[0001] The present invention relates to the technical field of rock physics modeling, and particularly relates to a solid-liquid decoupled fluid elastic impedance inversion method considering squeeze-extrusion effect. Background Art

[0002] In terms of rock physics considering the squeeze-extrusion effect, O’Connel (1974) and Budiansky (1976) pointed out that in addition to affecting the elasticity of the medium, fractures will also cause the squeeze-extrusion flow under the action of external forces, which is considered to be an important reason for the seismic wave velocity dispersion and energy attenuation in fluid-bearing rocks. In each frequency band, seismic waves are found to have strong velocity dispersion and energy attenuation characteristics. The fundamental factor dominating these characteristics is the non-uniformity of actual rocks at different scales.

[0003] Specifically analyze the squeeze-extrusion flow effect between pores and fractures, and relate this effect to the important parameters describing fractured media. The local non-uniformity inside rocks is widespread, mainly reflected in the differences in local porosity, non-uniform partial saturation, and multi-scale fractures. When elastic waves propagate in rocks, the local non-uniformity will cause the flow of fluid in pores at the mesoscopic scale, resulting in high-frequency dispersion and strong attenuation of elastic waves.

[0004] Reservoir fluid identification is one of the research hotspots in the fields of geophysics and oil and gas exploration. Reliable fluid identification results can improve the exploration success rate and reduce the input cost, which is of great significance for oil and gas reservoir exploration and reservoir evaluation. The elastic parameters of the medium can directly or indirectly participate in the construction process of fluid factors through algebraic operations, so as to realize the identification of reservoir pore fluid types. Prestack seismic inversion is an effective means to guide oil and gas identification.

[0005] In terms of elastic impedance inversion, elastic impedance is a generalization of acoustic impedance, expressed as a function of longitudinal wave velocity, shear wave velocity, density, and incident angle.

[0006] Among many influencing factors, the accuracy of the low-frequency model prior has the greatest impact on the reliability of inversion parameters and plays an important role in compensating the low-frequency components of conventional seismic inversion. Therefore, selecting a reliable initial model will improve the convergence accuracy and reliability of seismic inversion (Shin et al., 2008; Zong et al., 2016; Yin et al., 2016; Li et al., 2016b). However, reliable low-frequency parameter information is often difficult to obtain in actual data processing (new exploration areas with no wells or few wells), which will increase the difficulty of seismic inversion and reservoir characterization.

[0007] To address the problem of over-reliance on model accuracy in seismic inversion, experts at home and abroad have attempted to use band-limited seismic data and fuzzy parametric models to predict the ultra-low frequency information contained in model parameters. If the seismic attenuation effect is introduced into pure frequency-domain seismic inversion, a complex frequency-domain seismic inversion method will be generated, which will help alleviate the ill-posedness brought about by "band-limited inversion" (Sirgue et al., 2004; Brenders et al., 2007). Shin et al. (2008) studied the zero-frequency component of the attenuated seismic wavefield in complex frequency inversion and restored the long-wavelength velocity background of the subsurface model based on the full waveform inversion framework. Due to the problem of low penetration depth in this algorithm, complex Laplace-Fourier domain inversion introduced a series of frequency components (<5Hz) and attenuation coefficients, enhancing the input energy of parameters and penetration depth during the inversion process (Shin et al., 2009; Ha et al., 2012&2013; Hu et al., 2015). Seismic envelope inversion predicts the low-frequency velocity model in the envelope domain by constructing the target functional of the low-frequency envelopes of the synthetic signal and the actual signal (Wu et al., 2014; Luo et al., 2015&2016). Complex frequency-domain Laplace inversion and low-frequency envelope inversion are representatives of low-frequency inversion in different transform domains. Both aim to deeply explore the ultra-low frequency response in the original data, thereby improving the construction of low-frequency prior information for ill-posed problems and the reliability of prediction. Summary of the Invention

[0008] The object of the present invention is to provide a solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-out effect in view of the deficiencies in the prior art.

[0009] The technical solution is as follows:

[0010] A solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-out effect, comprising the following steps:

[0011] S1: Construct a solid-liquid decoupled fluid factor;

[0012] S2: Establish a relationship between the solid-liquid decoupled fluid factor and the elastic impedance;

[0013] S3: Invert the fluid factor based on Bayesian theory.

[0014] Further, the step S1 includes:

[0015] Define the longitudinal wave velocity V P and the transverse wave velocity V S as

[0016]

[0017]

[0018] Among them, β is the Biot coefficient, K d is the dry rock bulk modulus, M is the modulus parameter, μ is the shear modulus, ρ sat is the density of the saturated rock.

[0019] Furthermore, the step S1 includes:

[0020] The expression of the Gassmann fluid term f is:

[0021]

[0022] μ = μ d (4)

[0023] Among them, K sat , K s , K d , K f are the saturated fluid rock bulk modulus, the rock frame bulk modulus, the dry rock bulk modulus and the pore fluid equivalent bulk modulus, φ, μ, ρ are the equivalent porosity, the shear modulus and the density respectively, and μ d is the dry rock shear modulus.

[0024] Furthermore, the step S1 includes:

[0025] The Gassmann fluid factor is expressed as:

[0026]

[0027] Furthermore, the step S1 includes:

[0028] Assume:

[0029]

[0030] Therefore, the fluid factor can be expressed as:

[0031]

[0032] Compared with Equation (3), Equation (8) adds an S(ω) contribution term to characterize the jet flow:

[0033]

[0034]

[0035]

[0036]

[0037] where γ is the aspect ratio of the crack, η is the fluid viscosity, ε is the fracture density, ν can be the Poisson's ratio of the dry medium or the Poisson's ratio of the medium in the Biot phase theory, ω is the angular frequency, J is the Bessel equation, so Equation (8) can be expressed as:

[0038]

[0039]

[0040]

[0041] where K fs and f s are new fluid factors.

[0042] Furthermore, the step S2 includes:

[0043] The three-parameter reflection coefficient equation is:

[0044]

[0045] where R pp (θ) represents the reflection coefficient at an incident angle of θ, γ dry is the P-wave to S-wave velocity ratio in dry rock, γ sat is the P-wave to S-wave velocity ratio in saturated rock, f is the Gassmann fluid factor, μ is the shear modulus, and ρ is the density. According to f = G(φ)K f , the above equation is rewritten as

[0046]

[0047] Considering the squeeze-out effect, the expression of the reflection coefficient finally becomes

[0048]

[0049] where

[0050]

[0051]

[0052] S = 1 + S(ω)M (20)

[0053] According to the above reflection coefficient formula (17), combined with the relationship between the reflection coefficient and the elastic impedance EI, we have

[0054]

[0055] Therefore, the elastic impedance equation is

[0056] EI = (kfs)2a (μ s ) 2b (ρ) 2c (φ) 2d (S) 2e (22).

[0057] Furthermore, the step S2 includes:

[0058] Normalization processing results in

[0059]

[0060] wherein the parameter with subscript "0" is the initial value of each trace.

[0061]

[0062]

[0063]

[0064]

[0065]

[0066] According to the characteristics of pre-stack elastic impedance gather data, assuming there are 5 angle gathers and n sampling points, the elastic impedance can be expressed as

[0067]

[0068] To discuss the frequency-domain seismic inverse problem intuitively, assuming a specific frequency ω j the frequency-domain convolution model can be simplified to the following matrix equation

[0069] Y ωj = W ωj E ωj m (30)

[0070] In the formula, W ωj , E ωj represent the spectrum of the seismic wavelet and the Fourier forward operator respectively. Y ωj is the spectrum of the seismic data at a specific frequency ω j , and m is the model parameter to be inverted. The discrete Fourier forward operator E ωj is specifically

[0071]

[0072] In the formula, n is the number of time sampling points, τ nFor a discrete time series, assuming that the number of frequency selections is l, therefore, the matrix equation (30) expands to

[0073]

[0074] where Y is the seismic spectrum and E is the Fourier operator in the case of multiple frequencies.

[0075] Furthermore, the step S3 includes:[[]]

[0076] The posterior probability distribution of the model parameter m is

[0077] p(m|d) ∝ p(m)p(d|m) (33)

[0078] where d is the observed data, p(m|d) is the posterior probability, p(m) is the prior probability of the model parameter to be solved, and p(d|m) represents the similarity between the data obtained by forward modeling the inversion result again and the observed data.

[0079] Furthermore, the step S3 includes:[[]]

[0080] Assume that the noise in the observed data follows a Gaussian distribution with a mean of 0 and a covariance of X n then the likelihood function can be expressed as

[0081]

[0082] where G represents the forward operator of convolution.

[0083] Select the modified Cauchy distribution as the prior distribution of the reflection coefficient, and its mathematical expression is

[0084]

[0085] where is the variance of the time-domain model parameter, and m i is the discrete model parameter sequence.

[0086] Furthermore, the step S3 includes:[[]]

[0087] When the sampling points of the seismic data are N, first divide the model parameter vector into N groups according to the time sampling points, so that each group vector contains the P-wave reflection coefficient, S-wave reflection coefficient, density reflection coefficient, and quality factor reflection coefficient at this sampling point, that is

[0088]

[0089] where is the combination of 5 discrete model parameter sequences. Substitute the grouping situation into the modified Cauchy distribution to obtain the expression of the prior information of this method

[0090]

[0091] Among them, is the covariance of the model parameters Since seismic data is band-limited, it often leads to unstable inversion results and is greatly affected by noise. At this time, it can be improved by supplementing low-frequency information, that is, adding low-frequency constraints to the objective function:

[0092]

[0093] Among them, C is an operator with a diagonal of five integrals. ξ is a column vector including relative P-wave to S-wave velocity, relative density, and relative quality factor:

[0094]

[0095] In the above formula, K fs0 , μ S0 , ρ 0 , φ 0 , S 0 are the prior means of the five parameters.

[0096] Taking the improved sparse constraint and low-frequency constraint as the prior information of inversion, the final prior distribution is expressed as:

[0097] P(m) = P LFM (m)P mc (m) (40)

[0098] The final objective function is

[0099]

[0100] The beneficial effects of the present invention are:

[0101] Considering the influence of "local flow" at the pore scale in the elastic wave theory of Biot double-porous media, adding the squeeze-out flow effect between rock fractures, a new fluid indicator factor Kfs is constructed, and based on this, a five-parameter elastic impedance inversion equation is derived. The "solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-out effect" proposed by the present invention avoids the cumulative error of the conventional inversion that first obtains Vp, Vs, and ρ and then calculates the fluid factor from these three parameters. Theoretical analysis and verification of actual data show that this fluid factor has high accuracy and sensitivity for the detection of oil and gas in reservoirs. BRIEF DESCRIPTION OF THE DRAWINGS

[0102] Figure 1 A technical flowchart of a solid-liquid decoupled fluid factor elastic impedance inversion considering the squeeze-out effect

[0103] Figures 2(a)-(e) are the pre-stack seismic data required for five-parameter elastic impedance inversion, which are 12 degrees, 16 degrees, 21 degrees, 25 degrees, and 30 degrees respectively.

[0104] Figures 3(a)-(e) are the elastic impedance data at 5 angles required for five-parameter elastic impedance inversion, which are 12 degrees, 16 degrees, 21 degrees, 25 degrees, and 30 degrees respectively.

[0105] Figure 4 Inversion result of the Kfs fluid factor. Detailed implementation manners

[0106] To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below in combination with the detailed implementation manners and with reference to the accompanying drawings. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, the descriptions of well-known structures and technologies are omitted to avoid unnecessarily confusing the concepts of the present invention.

[0107] 1. Construction of the solid-liquid decoupled fluid factor

[0108] The main factors affecting seismic reflection characteristics are medium velocity and density. However, when the medium contains pores and fluids, the situation becomes more complicated. The influence of fluid saturation can be described by the Gassmann equation. According to the Biot-Gassmann theory, the longitudinal wave velocity V P and the shear wave velocity V S are defined as

[0109]

[0110]

[0111] where β is the Biot coefficient (its physical meaning is the ratio of the change in fluid volume to the change in rock volume when the pressure of water is constant), K d is the dry rock bulk modulus, M is the modulus parameter (its physical meaning is the pressure required to press the fluid into the rock without changing the formation volume), μ is the shear modulus, ρ sat is the density of the saturated rock.

[0112] The expression of the Gassmann fluid term f is:

[0113]

[0114] μ = μ d (4)

[0115] where K sat , K s , Kd , K f are the bulk modulus of the saturated fluid rock, the bulk modulus of the rock skeleton, the bulk modulus of the dry rock, and the equivalent bulk modulus of the pore fluid. φ, μ, and ρ are the equivalent porosity, shear modulus, and density respectively, and μ d is the shear modulus of the dry rock. The Gassmann fluid factor is expressed as:

[0116]

[0117] It can be reasonably assumed that:

[0118]

[0119] Therefore, the fluid factor can be expressed as:

[0120]

[0121] Tang Xiaoming (2011) proposed a new rock physics equation based on Biot theory, which maintains the basic characteristics and structure of Biot theory. Compared with Equation (3), Equation (8) adds an S(ω) contribution term to characterize the squirt flow:

[0122]

[0123]

[0124]

[0125]

[0126] where γ is the aspect ratio of the crack, η is the fluid viscosity, ε is the crack density, ν can be the Poisson's ratio of the dry medium or the Poisson's ratio of the medium in Biot's phase theory, ω is the angular frequency, J is the Bessel equation, so Equation (8) can be expressed as:

[0127]

[0128]

[0129]

[0130] where, K fs and f s are the new fluid factors. Compared with the conventional fluid factor types based on single-phase medium theory (Poisson's ratio σ, Lame parameter λ, and λρ, etc.), the solid-liquid decoupled fluid term K fsIt can better characterize the elastic influence of pore fluid on the rock medium, which is mainly reflected in that the fluid factor based on the bi-phase medium theory is much less affected by factors such as the rock solid skeleton than the fluid factor proposed based on the single-phase medium.

[0131] 2. Establishment of the relationship between the solid-liquid decoupled fluid factor and the elastic impedance

[0132] The three-parameter reflection coefficient equation of Russell (2011) is:

[0133]

[0134] Among them, R pp (θ) represents the reflection coefficient at the incident angle of θ, and γ dry is the P-S wave velocity ratio in dry rock, and γ sat is the P-S wave velocity ratio in saturated rock, f is the Gassmann fluid factor, μ is the shear modulus, and ρ is the density. According to f = G(φ)K f , the above formula is rewritten as

[0135]

[0136] Considering the squeeze-out effect, the expression of the reflection coefficient finally becomes

[0137]

[0138] Among them,

[0139]

[0140]

[0141] S = 1 + S(ω)M (20)

[0142] According to the above reflection coefficient formula (17), combined with the relationship between the reflection coefficient and the elastic impedance EI, we have

[0143]

[0144] Therefore, the elastic impedance equation is

[0145] EI = (kfs) 2a (μ s ) 2b (ρ) 2c (φ) 2d (S) 2e (22)

[0146] After normalization, we get

[0147]

[0148] Among them, the parameter with subscript "0" is the initial value of each pass.

[0149]

[0150]

[0151]

[0152]

[0153]

[0154] According to the characteristics of pre-stack elastic impedance gather data, assuming there are 5 angle gathers and n sampling points, the elastic impedance can be expressed as

[0155]

[0156] To discuss the seismic inverse problem in the frequency domain intuitively, under the assumption of a specific frequency ω j the frequency-domain convolution model can be simplified to the following matrix equation

[0157] Y ωj = W ωj E ωj m (30)

[0158] In the formula, W ωj , E ωj respectively represent the spectrum of the seismic wavelet and the Fourier forward operator. Y ωj is the seismic data spectrum at a specific frequency ω j , and m is the model parameter to be inverted. The discrete Fourier forward operator E ωj is specifically

[0159]

[0160] In the formula, n is the number of time sampling points, τ n is the discrete time series. Assuming the number of selected frequencies is l, therefore, the matrix equation (30) expands to

[0161]

[0162] In the formula, Y is the seismic spectrum, and E is the Fourier operator in the case of multiple frequencies.

[0163] 3. Fluid factor inversion based on Bayesian theory

[0164] Directly inverting the above equation is often an ill-posed problem. Therefore, it is necessary to add prior constraints to the objective function to make the inversion stable. Bayesian theory, based on known information, first assumes that the parameter to be solved follows a certain distribution, and then uses the likelihood function and the prior distribution to find the maximum a posteriori probability solution. The posterior probability distribution of the model parameter m is

[0165] p(m|d) ∝ p(m)p(d|m) (33)

[0166] where d is the observed data, p(m|d) is the posterior probability, p(m) is the prior probability of the model parameter to be solved, and p(d|m) represents the similarity between the data obtained by forward modeling the inversion result and the observed data.

[0167] Assume that the noise in the observed data follows a Gaussian distribution with a mean of 0 and a covariance of X n then the likelihood function can be expressed as

[0168]

[0169] where G represents the forward operator of convolution.

[0170] Select the modified Cauchy distribution as the prior distribution of the reflection coefficient, and its mathematical expression is

[0171]

[0172] where is the variance of the model parameter in the time domain, and m i is the discrete model parameter sequence. Substituting the above equation directly as prior information into Bayes' formula to obtain the objective function is reasonable in the logic of mathematics. However, to make the inversion result closer to the real situation, it is also necessary to consider the geophysical and geological significance of the parameter to be inverted, that is, the structural sparsity characteristics of the parameter to be inverted. The above equation is a sparse constraint on all parameters to be solved and does not consider the structural characteristics between the reflection coefficients of different parameters. Therefore, it is necessary to make certain improvements to the above formula. When the number of sampling points of the seismic data is N, first divide the model parameter vector into N groups according to the time sampling points, so that each group vector contains the P-wave reflection coefficient, S-wave reflection coefficient, density reflection coefficient, and quality factor reflection coefficient at this sampling point, that is

[0173]

[0174] where is the combination of 5 discrete model parameter sequences. Substituting the grouping situation into the modified Cauchy distribution to obtain the expression of the prior information of this method

[0175]

[0176] where For model parameters The covariance of. Since seismic data is band-limited, it often leads to unstable inversion results and is greatly affected by noise. At this time, it can be improved by supplementing low-frequency information, that is, adding low-frequency constraints to the objective function:

[0177]

[0178] Among them, C is an operator with five Integrals. ξ is a column vector, including relative P-wave to S-wave velocity, relative density, and relative quality factor:

[0179]

[0180] In the above formula, K fs0 , μ S0 , ρ 0 , φ 0 , S 0 Are the prior means of the five parameters.

[0181] Taking the improved sparse constraint and low-frequency constraint as the prior information of inversion, so the final prior distribution is expressed as:

[0182] P(m) = P LFM (m)P mc (m) (40)

[0183] The final objective function is

[0184]

[0185] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-jet effect, Characterized in that, It includes the following steps: S1: Construct a solid-liquid decoupled fluid factor; S2: Establish a relationship between the solid-liquid decoupled fluid factor and the elastic impedance; S3: Invert the fluid factor based on the Bayesian theory.

2. The solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-jet effect according to claim 1, Characterized in that, The step S1 includes: Define the longitudinal wave velocity V P and the shear wave velocity V S as Among them, β is the Biot coefficient, K d is the dry rock bulk modulus, M is the modulus parameter, μ is the shear modulus, ρ sat is the density of the saturated rock.

3. The solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-jet effect according to claim 2, Characterized in that, The step S1 includes: The expression of the Gassmann fluid term f is: μ = μ d (4) Among them, K sat , K s , K d , K f are the bulk modulus of the saturated fluid rock, the bulk modulus of the rock skeleton, the bulk modulus of the dry rock, and the equivalent bulk modulus of the pore fluid. φ, μ, and ρ are the equivalent porosity, shear modulus, and density respectively. μ d is the shear modulus of the dry rock.

4. The solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-jet effect according to claim 3, Characterized in that, The step S1 includes: The Gassmann fluid factor is expressed as:

5. The solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-jet effect according to claim 4, Characterized in that, The step S1 includes: Assume: Therefore, the fluid factor can be expressed as: Compared with equation (3), equation (8) adds an S(ω) contribution term to characterize the jet flow: Where γ is the aspect ratio of the crack, η is the fluid viscosity, ε is the crack density, ν can be the Poisson's ratio of the dry medium, or the Poisson's ratio of the medium in the Biot phase theory, ω is the angular frequency, J is the Bessel equation, so equation (8) can be expressed as: Among them, K fs and f s are new fluid factors.

6. The solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-jet effect according to claim 1, Characterized in that, The step S2 includes: The three-parameter reflection coefficient equation is: Among them, R pp (θ) represents the reflection coefficient at the incident angle of θ, γ dry is the P-wave to S-wave velocity ratio in dry rock, γ sat is the P-wave to S-wave velocity ratio in saturated rock, f is the Gassmann fluid factor, μ is the shear modulus, and ρ is the density. According to f = G(φ)K f , the above formula is rewritten as Considering the squeeze-jet effect, the expression of the reflection coefficient finally becomes Where, S = 1 + S(ω)M (20) According to the above reflection coefficient formula (17), combined with the relationship between the reflection coefficient and the elastic impedance EI, there is, Therefore, the elastic impedance equation is, EI = (kfs) 2a (μ s ) 2b (ρ) 2c (φ) 2d (S) 2e (22).

7. The solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-jet effect according to claim 1, Characterized in that, The step S2 includes: Normalization processing gives Where the parameters containing the subscript "0" are the initial values of each trace. According to the characteristics of the pre-stack elastic impedance gather data, assuming there are 5 angle gathers and n sampling points, the elastic impedance can be expressed as For an intuitive discussion of the frequency-domain seismic inverse problem, assuming a specific frequency ω j the frequency-domain convolution model can be simplified to the following matrix equation Y ωj = W ωj E ωj m(30) where W ωj , E ωj represent the spectrum of the seismic wavelet and the forward Fourier operator, respectively. Y ωj is the spectrum of seismic data at a specific frequency ω j , and m is the model parameter to be inverted. The discrete Fourier forward operator E ωj is specifically where n is the number of time sampling points, and τ n is a discrete time series. Assuming that the number of frequency selections is l, thus, the matrix equation (30) expands to In the formula, Y is the seismic spectrum, and E is the Fourier operator in the case of multiple frequencies.

8. The solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-jet effect according to claim 1, Characterized in that, The step S3 includes: The posterior probability distribution of the model parameter m is p(m|d) ∝ p(m)p(d|m) (33) Where d is the observed data, p(m|d) is the posterior probability, p(m) is the prior probability of the model parameter to be solved, and p(d|m) represents the similarity between the data obtained by forward modeling the inversion result again and the observed data.

9. The solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-jet effect according to claim 1, Characterized in that, The step S3 includes: Assume that the noise in the observed data follows a Gaussian distribution with a mean of 0 and a covariance of X n , then the likelihood function can be expressed as Where G represents the forward operator of convolution. Select the modified Cauchy distribution as the prior distribution of the reflection coefficient, and its mathematical expression is Among them, is the variance of the time-domain model parameter, m i is the discrete model parameter sequence.

10. A solid-liquid decoupled fluid elastic impedance inversion method considering the squeeze-out effect according to claim 1, characterized in that the step S3 includes: When the sampling points of the seismic data are N, first divide the model parameter vector into N groups according to the time sampling points, so that each group of vectors contains the P-wave reflection coefficient, S-wave reflection coefficient, density reflection coefficient and quality factor reflection coefficient at the sampling point, that is Among them, is a combination of 5 discrete model parameter sequences. Substituting the grouping situation into the modified Cauchy distribution to obtain the expression of the prior information of this method Among them, is the covariance of the model parameters Since seismic data is band-limited, it often leads to unstable inversion results and is greatly affected by noise. At this time, it can be improved by supplementing low-frequency information, that is, adding low-frequency constraints to the objective function: where C is an operator with a diagonal of five integrals. ξ is a column vector including relative P- to S-wave velocity, relative density, and relative quality factor: In the above formula, K fs0 , μ S0 , ρ 0 , φ 0 , S 0 are the prior means of the five parameters. Taking the improved sparse constraint and low-frequency constraint as the prior information of the inversion, so the final prior distribution is expressed as: P(m) = P LFM (m)P mc (m) (40) The final objective function is

Images