Small-scale fracture development region characterization method and system based on high-precision speed model

Abstract

The invention relates to the technical field of oil-gas exploration, in particular to a small-scale fracture development area characterization method and system based on a high-precision velocity model, and the method comprises the steps: correcting a seismic interval velocity function of a target area through logging data, obtaining interval velocity information through the interval velocity function, and building a high-precision velocity model; obtaining adaptive normal illumination data by using a connectivity attribute body based on multi-parameter optimization, and converting the adaptive normal illumination data of a time domain into adaptive normal illumination data of a depth domain by using a high-precision speed model; the method comprises the following steps: establishing an orientation guidance field by using fracture orientation and fracture density, obtaining initial fracture grid models of different wells by using the orientation guidance field and the fracture density optimized layer by layer in logging, and outputting a dynamic small-scale fracture grid model through model weight distribution, optimal weight value screening and high-density grid connection. And determining a fracture development area according to the density and orientation of the small-scale fracture, wherein the dynamic small-scale fracture grid model is used for representing the fracture development area and spatial distribution characteristics. According to the method, the spatial distribution rule of the small-scale crack dense development area can be quickly and accurately judged.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas exploration, and in particular to a method and system for characterizing a small-scale fracture development zone based on a high-precision velocity model. Background Art

[0002] Fractures are widely developed in the crustal medium. They are formed under the action of crustal stress after a long period of geological tectonic movement. Fractures themselves can serve as oil and gas storage space, as well as important channels for oil and gas migration, controlling the formation and distribution of oil and gas reservoirs. Therefore, fractures are important seismic parameters for oil and gas reservoir exploration and development, and many scholars have used a variety of fracture modeling methods to study them.

[0003] Conventional fracture modeling methods: First, multi-parameter analysis is performed on pre-stack seismic data, and the problems of near-channel multiple wave interference noise and limited bandwidth in three-dimensional seismic data are systematically analyzed, and the optimization processing of relevant seismic gathers is further carried out. Then, the fracture sensitivity attributes are extracted from the denoised seismic data. Further, through comparative analysis of different parameters such as incident angles and azimuth angles, the optimal fracture sensitivity attribute data in the time domain and the normal illumination data in the time domain are calculated in turn, so as to roughly determine the distribution range of the fracture development zone. Conventional fracture modeling methods can identify the laws of reservoir fracture development zones in areas with large-scale fracture development by using discrete fracture grid modeling technology. However, in areas with small-scale fracture development, the application effect is poor. Summary of the invention

[0004] To this end, the present invention provides a method and system for characterizing small-scale fracture development zones based on a high-precision velocity model to solve the problem of unsatisfactory application effect of small-scale fracture modeling in oil and gas exploration.

[0005] According to the design scheme provided by the present invention, on the one hand, a method for characterizing a small-scale fracture development zone based on a high-precision velocity model is provided, comprising:

[0006] Obtain pre-stack OVT domain seismic gather data and perform noise analysis;

[0007] AVO fitting compensation gather optimization processing, wherein the AVO fitting compensation gather optimization processing includes fitting a cubic polynomial, removing noise and target gather energy compensation;

[0008] Use logging data to correct the seismic layer velocity function of the target area, and use the layer velocity function to obtain layer velocity information to establish a high-precision velocity model;

[0009] The time domain adaptive normal lighting data is obtained by using the connectivity attribute body based on multi-parameter optimization, and the time domain adaptive normal lighting data is converted into depth domain adaptive normal lighting data by using a high-precision velocity model;

[0010] An orientation guidance field is established using the fracture orientation and fracture density, and an initial small-scale fracture grid model and a dynamic small-scale fracture grid model are established using the orientation guidance field and fracture density corrected layer by layer by a neural network. The dynamic small-scale fracture grid model is used to characterize fracture development areas and spatial distribution characteristics;

[0011] The fracture development zone is determined according to the small-scale fracture density and fracture orientation and based on the dynamic small-scale fracture grid model to characterize the small-scale fracture development area.

[0012] As a characterization method of small-scale fracture development area based on high-precision velocity model in the present invention, further, the fitting cubic polynomial in the AVO fitting compensation gather optimization process is expressed as T=P+Gx+P*Gx 2 +Cx 3 , where T is the reflection coefficient, x is the incident angle, P is the intercept, G is the gradient, P*G is the AVO indicator factor, and C is the curvature.

[0013] As a method for characterizing a small-scale fracture development zone based on a high-precision velocity model of the present invention, further, using well logging data to correct the seismic layer velocity data of the target area, it includes:

[0014] Extract the layer velocity data at the target well point, perform correlation analysis and velocity correction on the seismic layer velocity and the layer velocity at the well point to obtain the corrected seismic layer velocity data;

[0015] The corrected seismic layer velocity data are processed by thinning, interpolation and smoothing in turn to obtain a high-precision velocity model.

[0016] As a method for characterizing a small-scale fracture development zone based on a high-precision velocity model of the present invention, further, correlation analysis and velocity correction are performed on the seismic layer velocity and the layer velocity at the well point, including:

[0017] The cross-correlation analysis function is used to analyze the cross-correlation between the seismic layer velocity and the layer velocity at the well point, so as to obtain the difference between the seismic layer velocity and the layer velocity at the well point in different strata;

[0018] Based on the difference between the seismic layer velocity and the well point layer velocity of different strata, a correction function is constructed by using the layer velocity value of the well and the seismic layer velocity value, so as to correct the seismic layer velocity value by using the correction function;

[0019] The corrected seismic layer velocities are sequentially subjected to three-dimensional high-density gridding processing to obtain a high-precision velocity model through thinning, interpolation and smoothing.

[0020] As a method for characterizing a small-scale fracture development zone based on a high-precision velocity model of the present invention, further, the process of converting the adaptive normal illumination data in the time domain into the adaptive normal illumination data in the depth domain using the high-precision velocity model is expressed as: F (h) = V vol ΔγC F (t), where C F (t) is the adaptive normal illumination data in the time domain, Δγ is the adaptive factor, C F (h) is the adaptive normal lighting data in the depth domain, V vol It is the seismic layer velocity function in the high-precision velocity model.

[0021] As a method for characterizing a small-scale fracture development zone based on a high-precision velocity model of the present invention, further, under the constraint condition of the azimuth guidance field of different wells, the azimuth guidance field and fracture density based on layer-by-layer optimization of a neural network unit are used to obtain an initial fracture grid model and a dynamic small-scale fracture grid model, comprising:

[0022] The crack density function and crack orientation function are constructed based on the crack circular discrete space function, the crack circular space radius and the vertical and horizontal coordinates;

[0023] The fracture density function is normalized according to the maximum and minimum values ​​of the fracture density, the normalized fracture density and fracture orientation are subjected to high-density gridding, and the orientation guidance field is obtained;

[0024] An azimuth guidance field is established according to the fracture orientation and fracture density. The azimuth guidance fields of different wells are used to correct the azimuth guidance fields in the neural network unit layer by layer. The corrected azimuth guidance field and fracture density are used to establish an initial small-scale fracture grid model of multiple wells. Furthermore, model weight allocation, optimal weight value screening and high-density grid connection are performed in turn for the initial small-scale fracture grid model of multiple wells to obtain a dynamic small-scale fracture grid model.

[0025] As a method for characterizing a small-scale fracture development zone based on a high-precision velocity model of the present invention, further, the fracture development zone is determined according to the density and orientation of the small-scale fractures, comprising:

[0026] The fracture density and fracture orientation are linearly regressed, and the scope and law of the small-scale fracture development zone are determined by combining the planar distribution characteristics of the fracture development zone in the dynamic small-scale fracture grid model.

[0027] On the other hand, the present invention also provides a small-scale fracture development zone characterization system based on a high-precision velocity model, comprising: a data acquisition module, a data optimization module, a velocity model acquisition module, a data conversion module, a target model establishment deep learning module and a regional characterization module, wherein:

[0028] Data acquisition module, used to obtain pre-stack OVT domain seismic gather data and perform noise analysis;

[0029] A data optimization module, used for AVO fitting compensation gather optimization processing, wherein the AVO fitting compensation gather optimization processing includes fitting a cubic polynomial, removing noise and target gather energy compensation;

[0030] The velocity model acquisition module is used to correct the seismic layer velocity function of the target area using the logging data, and obtain the layer velocity information using the layer velocity function to establish a high-precision velocity model;

[0031] A data conversion module, used for obtaining time-domain adaptive normal lighting data by using a multi-parameter optimized connectivity attribute body, and converting the time-domain adaptive normal lighting data into depth-domain adaptive normal lighting data by using a high-precision velocity model;

[0032] The target model establishes a deep learning module, which is used to establish an orientation guidance field using fracture orientation and fracture density, and to correct the orientation guidance field in the neural network unit layer by layer using the orientation guidance fields of different wells, and to establish an initial small-scale fracture grid model of multiple wells using the corrected orientation guidance field and fracture density. Furthermore, the initial small-scale fracture grid model of multiple wells is sequentially assigned model weights, and the optimal weight values ​​and high-density grid connections are screened to obtain a dynamic small-scale fracture grid model, which is used to characterize fracture development areas and spatial distribution characteristics.

[0033] The regional characterization module is used to determine the fracture development area according to the density and orientation of small-scale fractures and based on the dynamic small-scale fracture grid model to characterize the small-scale fracture development area.

[0034] Beneficial effects of the present invention:

[0035] The present invention corrects seismic layer velocity data through logging data, establishes a high-precision velocity model, and provides high-precision velocity field information for deep-domain seismic data; performs (CF) connectivity calculation on seismic data with near-channel multiple wave interference removed, obtains connectivity attribute body, obtains adaptive normal illumination data in time domain through screening of parameters such as incident angle and azimuth, and converts it into adaptive normal illumination data in depth domain using a high-precision velocity model, which can accurately reflect the distribution law of small-scale fracture development areas; establishes an azimuth guidance field through fracture azimuth and fracture density, and under the constraint condition of the azimuth guidance field of different wells, uses the azimuth guidance fields of different wells to correct the azimuth guidance fields in neural network units layer by layer, and uses the corrected azimuth guidance fields and fracture densities to establish initial small-scale fracture development areas for multiple wells in turn. The dynamic small-scale fracture grid model is a novel fracture grid model. Furthermore, the model weights are allocated, the optimal weight values ​​are screened, and the high-density grid connections are performed in turn for the initial small-scale fracture grid models of multiple wells to obtain a dynamic small-scale fracture grid model. The dynamic small-scale fracture grid model can better reflect the development law of the small-scale fracture system and provide an important basis for the subsequent well site deployment. On the basis of studying the planar distribution characteristics of the first type of fracture development zone of the dynamic small-scale grid model, a linear regression analysis is carried out on the fracture density and fracture orientation, focusing on the classification rules and development zone ranges of the regional small-scale fracture system, so as to quickly and accurately determine the spatial distribution law of the small-scale fracture dense development zone, which is convenient for the subsequent reasonable deployment of wells and oil field exploitation, and has a good application prospect in the field of oil field exploration. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a schematic diagram of the process of characterizing the small-scale fracture development zone based on the high-precision velocity model in the embodiment;

[0037] Figure 2 This is the AVO fitting compensation gather before optimization processing in the embodiment;

[0038] Figure 3 The AVO fitting compensation gathers in the embodiment are optimized;

[0039] Figure 4 A high-precision three-dimensional velocity model in the embodiment;

[0040] Figure 5 is a depth domain adaptive normal lighting plane map in the embodiment;

[0041] Figure 6 It is the crack orientation plane distribution diagram-A in the embodiment;

[0042] Figure 7 It is the crack orientation plane distribution diagram-B in the embodiment;

[0043] Figure 8 The dynamic small-scale fracture grid model-A in the embodiment;

[0044] Fig. 9 It is the dynamic small-scale fracture grid model-B in the embodiment;

[0045] Fig.10 It is the three-dimensional dynamic small-scale crack grid model-B in the embodiment;

[0046] Fig.11 It is the linear regression analysis of the fracture development zone in the embodiment;

[0047] Fig.12 This is the dynamic small-scale fracture grid model in the embodiment - the layout diagram of the first-level fracture development zone. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the present invention clearer and more understandable, the present invention is further described in detail below in conjunction with the accompanying drawings and technical solutions.

[0049] In view of the fact that the conventional fracture development zone characterization method is insufficient in time domain data accuracy in the application of oil and gas exploration, it is difficult to effectively identify the area where small-scale fractures are more developed, and it is easy to add interference information of seismic data while identifying the small-scale fracture development area, the embodiment of the present invention, see Figure 1 As shown, a method for characterizing small-scale fracture development zones based on a high-precision velocity model is provided, comprising:

[0050] P001. Obtain pre-stack OVT domain seismic gather data and perform noise analysis.

[0051] The OVT domain prestack gather data with the three-dimensional prestack seismic data of the target oil and gas block as input is used to conduct multi-parameter analysis on the prestack gather data. Noise analysis is conducted on the original seismic gather data. Multiple wave interference is not conducive to the subsequent study of fracture development areas, and optimization processing is performed by removing near-channel multiple waves.

[0052] P002. AVO fitting compensation gather optimization processing The AVO fitting compensation gather optimization processing includes fitting a cubic polynomial, removing noise and compensating the target gather energy.

[0053] Through application in a certain work area, taking the three-dimensional prestack seismic data of the work area as an example, the formation dip in the study area is relatively gentle and small-scale fractures are relatively developed. In view of the near-channel multiple wave interference problem in the study area, the AVO fitting compensation processing is first carried out on the OVT domain prestack gather data, which specifically includes the establishment of a cubic fitting equation and energy compensation of the target gather to separate the near-channel multiple wave interference noise and the effective signal. The specific steps can be described as follows:

[0054] The specific steps are as follows:

[0055] The first step is to input the pre-stack OVT seismic data. Through the gather analysis, it can be seen that the multiple wave interference is relatively serious. It is urgent to establish a cubic fitting equation for the amplitude changing with the offset distance. The formula has a parabolic feature, in which the terms to be calculated are the incident angle x, P (intercept), G (gradient), P*G (AVO indicator factor) and C (curvature). The specific formula is as follows:

[0056] T=P+Gx+P*Gx 2 +Cx 3 (1)

[0057]

[0058] From formula (1) and formula (2), we know that T is the reflection coefficient, P, G, P*G and C are the coefficients of the equation, V P 、V S , D ρ are the velocity and density of longitudinal wave impedance, transverse wave impedance, ΔV P , ΔV S and ΔD ρ are the velocity change rate of longitudinal wave impedance, velocity change rate of shear wave impedance and density change rate respectively, and δ is the incident angle. Define Q as the forward function, and convolve the reflection coefficient with the Ricker wavelet. The specific expression is as follows:

[0059] Q=T*W (3)

[0060] According to formula (3), Q is the forward function, T is the reflection coefficient, W is the Ricker wavelet, and * is the convolution operator symbol.

[0061] In the second step, according to the error function calculation, the effective wave value with smaller error is retained, and the near-channel multiple waves with larger error are eliminated. The AVO fitting compensation expression is as follows:

[0062]

[0063] According to formula (4), Z is the fitting function, S is the compensation factor, f is the measured function, Q is the forward function, [] E is transposed.

[0064] pass Figure 2 and Figure 3 By comparison, it can be seen that the near-channel multiple wave removal effect is obvious, the continuity of the event axis is improved, and the signal-to-noise ratio is significantly improved, which provides an important data source for the later small-scale crack characterization.

[0065] P003. Use logging data to correct the seismic layer velocity function in the target area, and use the layer velocity function to obtain layer velocity information to establish a high-precision velocity model.

[0066] Specifically, the target area seismic layer velocity data is corrected using well logging data, which can be designed to include:

[0067] Extract the layer velocity data at the target well point, perform correlation analysis and velocity correction on the seismic layer velocity and the layer velocity at the well point to obtain the corrected seismic layer velocity data;

[0068] The corrected seismic layer velocity data are processed by thinning, interpolation and smoothing in turn to obtain a high-precision velocity model.

[0069] Since the traditional fracture grid modeling method cannot effectively reflect the dense spatial distribution pattern of fractures in areas with small formation inclination angles and developed small-scale fractures, in the embodiment of this case, the optimal connectivity attribute body is obtained by adjusting multiple parameters step by step, and the adaptive normal illumination data in the time domain is calculated for the optimal connectivity attribute. On this basis, it is converted into adaptive normal illumination data in the depth domain through a high-precision velocity model.

[0070] The correlation analysis and velocity correction of the seismic layer velocity and the layer velocity at the well point may include:

[0071] The cross-correlation analysis function is used to calculate the cross-correlation between the seismic layer velocity and the layer velocity at the well point, so as to obtain the difference between the seismic layer velocity and the layer velocity at the well point in different strata;

[0072] Based on the difference between the seismic layer velocity and the well point layer velocity of different strata, a correction function is constructed by using the layer velocity value of the well and the seismic layer velocity value, so as to correct the seismic layer velocity value by using the correction function;

[0073] The corrected seismic layer velocities are sequentially subjected to three-dimensional high-density gridding processing to obtain a high-precision velocity model through thinning, interpolation and smoothing.

[0074] The establishment of a high-precision velocity model can be summarized as the following four steps. The specific steps are as follows:

[0075] The first step is to define the target layer segment and use the stratigraphic data to establish different stratigraphic structural surfaces.

[0076] The second step is to define the logging layer data to correct the structural planes of different formations and generate an accurate geological framework model.

[0077] The third step is to define the earthquake root mean square velocity V RMS , the seismic layer velocity is V vol , the interval velocity at the well point is W vol , using the seismic root mean square velocity to convert to seismic layer velocity, the expression is as follows:

[0078]

[0079] According to formula (5), V SMR is the root mean square velocity, V vol is the seismic layer velocity, t top is the time value of the top interface of the formation, t down is the bottom interface time value of the formation, is the formation inclination.

[0080] The seismic layer velocity and the layer velocity at the well point are cross-correlated and the difference between the seismic layer velocity and the layer velocity at the well point in different strata is analyzed. The specific expression is as follows:

[0081]

[0082] According to formula (6), α is the correlation factor, t is the time value, υ is the time delay value, and W vol is the layer velocity function at the well point, V vol is the seismic layer velocity function.

[0083] Furthermore, the correction function is constructed using the well layer velocity value and the seismic layer velocity value, as described in Table 1, to obtain the seismic layer velocity value with higher accuracy. The correction function expression is as follows:

[0084]

[0085] According to formula (7), t is the time value, W vol is the layer velocity function at the well point, V vol is the seismic layer velocity function, L2 is the L2 norm function, that is, the correction function. The smaller the correction function is, the closer the well layer velocity is to the seismic layer velocity, which means the seismic layer velocity accuracy is higher.

[0086] Table 1 Velocity parameter information of seismic and well logging

[0087]

[0088] The fourth step is to analyze the characteristics of small-scale fractures in the study area, perform three-dimensional high-density gridding on the corrected layer velocity, carry out thinning, interpolation and smoothing, and output a three-dimensional high-precision velocity model, such as Figure 4 shown.

[0089] P004. Adaptive normal lighting data in the time domain is obtained by using a connectivity attribute body based on multi-parameter optimization, and the adaptive normal lighting data in the time domain is converted into adaptive normal lighting data in the depth domain by using a high-precision velocity model.

[0090] The connectivity (CF) attribute volume is calculated for the seismic data volume after multiple wave removal, and the best connectivity (CF) attribute volume is obtained through multi-parameter comparative analysis, including different incident angles, azimuths and frequency information. The adaptive normal illumination data in the time domain is calculated for the best connectivity attribute volume, and the time domain data is converted into depth domain data based on the high-precision velocity model, such as Figure 5 As shown, the corrected seismic layer velocity function V is defined as vol , the specific formula is as follows:

[0091] C F (h) = V vol ΔγC F (t) (8)

[0092] According to formula (8), C F (t) is the adaptive normal illumination data in the time domain after filtering by the incident angle, azimuth and frequency information. It is worth noting that the adaptive normal illumination data in the time domain adds an adaptive factor Δγ, C F (h) is the adaptive normal lighting data in the depth domain, V vol Corrected seismic layer velocity function.

[0093] P005. Use the fracture orientation and fracture density to establish an orientation guidance field. Take the orientation guidance fields of different wells as constraints. Use the orientation guidance fields and fracture density corrected layer by layer by the neural network unit to establish the initial small-scale fracture grid model of multiple wells. Obtain a dynamic small-scale fracture grid model through model weight allocation, screening of optimal weight values, and high-density grid connection processing. The dynamic small-scale fracture grid model is used to characterize the fracture development area and spatial distribution characteristics. Specifically, it can be designed to include:

[0094] Step 1: Construct the crack density function and crack orientation function based on the crack circular discrete space function, crack circular space radius and vertical and horizontal coordinates;

[0095] Step 2: Normalize the fracture density function according to the maximum and minimum values ​​of the fracture density, perform high-density grid processing on the normalized fracture density and fracture orientation, and obtain the orientation guidance field;

[0096] Step 3: Discretize the azimuth guidance field function and the fracture density function, establish a multi-layer neural network unit, use the azimuth guidance fields of different wells to correct the azimuth guidance fields in the neural network unit layer by layer, and use the corrected azimuth guidance fields and fracture density to obtain the initial small-scale fracture grid model of multiple wells.

[0097] The fourth step is to analyze the neural network units of different layers in multiple wells, assign model weights to the initial small-scale fracture grid models of multiple wells, screen the optimal weight values ​​and high-density grid connections, and output a dynamic small-scale fracture grid model suitable for the study area.

[0098] Among them, in the first step, the function expressions of crack density and crack orientation can be expressed as follows:

[0099]

[0100] According to formula (9) and formula (10), P(x, y) is the crack density, θ(x, y) is the crack orientation, S() is the crack circular discrete space function, (x, y) is the horizontal and vertical coordinates of the crack circular space, R is the radius of the crack circular space, D is the crack density coefficient, and θ is the crack orientation coefficient.

[0101] In the second step, the crack density is normalized, and the function expression is as follows:

[0102]

[0103] According to formula (11), is the normalized crack density, (x, y) is the spatial horizontal and vertical coordinates of the crack, is the normalized crack density function, max() is the maximum crack density, and min() is the maximum crack density.

[0104] The crack density and crack orientation are processed by high-density gridding to obtain the expression of the orientation guidance field ω:

[0105]

[0106] According to formula (12), ω(x, y) is the azimuth guidance field, (x, y) is the spatial horizontal and vertical coordinates of the crack, and ω is the crack density function. is the normalized fracture density function, θ is the normalized fracture orientation function, λ is the anisotropy coefficient, is the normalized crack density, θ(x, y) is the crack orientation, * is the convolution operator symbol, and Q is a constant.

[0107] In the third step, the azimuth guidance field and crack density are discretized, and the azimuth guidance field function expression is as follows:

[0108]

[0109] Among them, ω RZ (x, y) is the azimuth guidance field at different times, R is the lateral offset increment, Z is the depth increment, [] is the discretization matrix, ω 11(x,y),ω 12 (x,y),…,ω 1n (x, y) is the azimuth guidance field of the first layer with different offsets; ω 21 (x,y),ω 22 (x,y),…,ω 2n (x,y) is the azimuth guidance field ω of the second layer with different offsets n1 (x,y),ω n2 (x,y)…ω nm (x, y) is the azimuth guidance field of the same lateral offset, representing the azimuth guidance field of different offsets at the nth layer. R∈[1,n], Z∈[1,m], n and m are the maximum values ​​of the lateral offset and depth, respectively.

[0110] The crack density at different depths is defined as The discretization function expression of crack density is:

[0111]

[0112] in, is the discretized fracture density at different depths, is the discretized crack density function.

[0113] Using ω RZ The (x,y) azimuth guidance field matrix variable is used to establish neural network units at different layers, and the azimuth guidance fields of different wells are used to correct the azimuth guidance fields in the neural network units layer by layer. 11 (x,y),ω 12 (x,y)…ω 1m (x,y) form the first layer of neural network; then, ω 21 (x,y),ω 22 (x,y)…ω 2m (x, y) form the second layer of neural network units; and so on, ω n1 (x,y),ω n2 (x,y)…ω nm (x, y) form the nth layer of neural network units. Use the azimuth guidance fields of wells MB1, MB2, MB3, MB4 and MB5 to correct the azimuth guidance fields of neural network units layer by layer, and obtain the azimuth guidance fields ω of wells MB1, MB2, MB3, MB4 and MB5. RZ-MB1 (x,y),ω RZ-MB2 (x,y),ω RZ-MB3 (x,y),ω RZ-MB4 (x,y) and ω RZ-MB5 (x,y).

[0114] The azimuth guidance fields of different wells are respectively related to the fracture density at different depths. Fracture modeling was performed to obtain the initial small-scale fracture grid models of wells MB1, MB2, MB3, MB4 and MB5.

[0115] In the fourth step, the neural network units of different layers of multiple wells are analyzed, and the model weight allocation, optimal weight value screening and high-density grid connection are performed on the initial small-scale fracture grid model of multiple wells in turn. The weight allocation table of the initial small-scale fracture model of multiple wells is as follows:

[0116] Table 2 Initial small-scale crack model weight distribution table

[0117]

[0118] The model weight coefficients of wells MB1, MB2, MB3, MB4 and MB5 are shown in Table 2. The optimal weight value is 0.27, which is the initial small-scale fracture grid model of well MB5. A high-density grid connection calculation is performed on the initial small-scale fracture model to obtain a dynamic small-scale fracture grid model.

[0119] Figure 6 and 8 It is a fracture orientation and small-scale discrete fracture grid model obtained by conventional small-scale fracture characterization method. Figure 7 , Fig. 9 and 10 The fracture orientation, small-scale fracture grid model and three-dimensional small-scale fracture grid model obtained by this scheme. Among them, the conventional small-scale fracture characterization method can be mainly summarized as including the following steps:

[0120] T001: Input OVT domain pre-stack gather data, such as Figure 2 As shown, multi-parameter analysis of pre-stack seismic gathers is carried out. Noise analysis is carried out on the original seismic data;

[0121] T002: Aiming at the near-channel multiple wave interference problem in the study area, AVO fitting compensation processing is carried out on the OVT domain pre-stack gather data, which specifically includes the establishment of a cubic fitting equation and energy compensation of the target gather to separate the near-channel multiple wave interference noise and the effective signal.

[0122] T003: Calculate the connectivity CF attribute volume for the seismic data volume after removing multiple waves, and obtain the best connectivity CF attribute volume through multi-parameter comparative analysis, which includes different incident angles, azimuths and frequency information. Furthermore, calculate the normal illumination data in the time domain for the best connectivity attribute volume.

[0123] T004: Establish an orientation guidance field based on the fracture density and fracture orientation, and use the orientation guidance field to further establish a small-scale discrete fracture grid model to determine the small-scale fracture development area.

[0124] Depend on Figure 6 and Figure 7 By comparison, the small-scale fracture system in this scheme is more delicately described. Figure 8 , Fig. 9 and Fig.10 By comparison, the three-dimensional dynamic small-scale fracture grid model in this scheme can better characterize the fracture development area and spatial distribution characteristics.

[0125] P006. Perform linear regression on the fracture density and fracture orientation to determine the level 1 fracture development area in the intersection diagram.

[0126] Among them, the specific steps of linear regression analysis can be described as follows:

[0127] L=λX1 2 +δX2+Z (15)

[0128] In formula (15), L is the normalized regression equation function of the fracture density and fracture orientation, X1 is the fracture density term, X2 is the fracture orientation term, λ is the fracture density coefficient, δ is the isotropy coefficient, and Z is a constant.

[0129] Depend on Fig.11 It can be seen that the vertical axis represents the crack density, the horizontal axis represents the crack orientation, and the blue scatter plot has good linear characteristics. Therefore, the blue scatter area is divided into Class 1 crack development area.

[0130] P007. Determine the scope and law of small-scale fracture development zone by combining the plane distribution characteristics of fracture development zone in dynamic small-scale fracture grid model.

[0131] Based on the comprehensive analysis of the linear regression analysis results of the fracture development zone and the planar distribution characteristics of the type 1 fracture development zone of the dynamic small-scale grid model, the scope and law of the small-scale fracture development zone are basically determined.

[0132] Fig.11 Linear regression analysis of fracture development area, Fig.12 is the plane distribution diagram of the first-level fracture development area of ​​the dynamic small-scale grid model, Fig.11 The blue scatter plot in the middle represents the first-level fracture development area. Fig.12 The middle red range also represents the level 1 fracture development zone, which can ultimately determine the scope of the small-scale fracture development zone in the study area.

[0133] Furthermore, based on the above method, an embodiment of the present invention also provides a small-scale fracture development zone characterization system based on a high-precision velocity model, comprising: a data acquisition module, a data optimization module, a velocity model acquisition module, a data conversion module, a target model establishment module and a regional characterization module, wherein:

[0134] Data acquisition module, used to obtain pre-stack OVT domain seismic gather data and perform noise analysis;

[0135] A data optimization module, used for AVO fitting compensation gather optimization processing, wherein the AVO fitting compensation gather optimization processing includes fitting a cubic polynomial, removing noise and target gather energy compensation;

[0136] The velocity model acquisition module is used to correct the seismic layer velocity function of the target area using the logging data, and obtain the layer velocity information using the layer velocity function to establish a high-precision velocity model;

[0137] A data conversion module, used for obtaining time-domain adaptive normal lighting data by using a multi-parameter optimized connectivity attribute body, and converting the time-domain adaptive normal lighting data into depth-domain adaptive normal lighting data by using a high-precision velocity model;

[0138] The target model establishes a deep learning module, which is used to establish an orientation guidance field using fracture orientation and fracture density, and to correct the orientation guidance field in the neural network unit layer by layer using the orientation guidance fields of different wells, and to establish an initial small-scale fracture grid model of multiple wells using the corrected orientation guidance field and fracture density. Furthermore, the initial small-scale fracture grid model of multiple wells is sequentially assigned model weights, and the optimal weight values ​​and high-density grid connections are screened to obtain a dynamic small-scale fracture grid model, which is used to characterize fracture development areas and spatial distribution characteristics.

[0139] The regional characterization module is used to determine the fracture development area according to the density and orientation of small-scale fractures and based on the dynamic small-scale fracture grid model to characterize the small-scale fracture development area.

[0140] Unless otherwise specifically stated, the relative steps, numerical expressions and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present invention.

[0141] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0142] The units and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A person of ordinary skill in the art may use different methods to implement the described functions for each specific application, but such implementation is not considered to be beyond the scope of the present invention.

[0143] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits, and accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of software function modules. The present invention is not limited to any specific form of combination of hardware and software.

[0144] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for characterizing small-scale fracture development zones based on a high-precision velocity model, characterized in that: Include: Obtain pre-stack OVT domain seismic gather data and perform noise analysis; AVO fitting compensation gather optimization processing, wherein the AVO fitting compensation gather optimization processing includes fitting a cubic polynomial, removing noise and target gather energy compensation; Use logging data to correct the seismic layer velocity function of the target area, and use the layer velocity function to obtain layer velocity information to establish a high-precision velocity model; The time domain adaptive normal lighting data is obtained by using the connectivity attribute body based on multi-parameter optimization, and the time domain adaptive normal lighting data is converted into depth domain adaptive normal lighting data by using a high-precision velocity model; An orientation guidance field is established using fracture orientation and fracture density, and the orientation guidance fields in the neural network unit are corrected layer by layer using the orientation guidance fields of different wells. An initial small-scale fracture grid model is established using the corrected orientation guidance field and fracture density, and model weight allocation, optimal weight value screening and high-density grid connection are performed in sequence for the initial small-scale fracture grid model to obtain a dynamic small-scale fracture grid model, which is used to characterize fracture development areas and spatial distribution characteristics; The fracture development zone is determined according to the density and orientation of small-scale fractures and based on the dynamic small-scale fracture grid model to characterize the small-scale fracture development area.

2. The method for characterizing small-scale fracture development zones based on a high-precision velocity model according to claim 1 is characterized in that: The fitting cubic polynomial in the AVO fitting compensation gather optimization process is expressed as T = P + Gx + P*Gx 2 +Cx 3 , where T is the reflection coefficient, x is the incident angle, P is the intercept, G is the gradient, P*G is the AVO indicator factor, and C is the curvature.

3. The method for characterizing small-scale fracture development zones based on a high-precision velocity model according to claim 1 is characterized in that: Use logging data to correct the seismic layer velocity function in the target area, including: Extract the layer velocity data at the target well point, perform correlation analysis and velocity correction on the seismic layer velocity and the layer velocity at the well point to obtain the corrected seismic layer velocity data; The corrected seismic layer velocity data are processed by thinning, interpolation and smoothing in turn to obtain a high-precision velocity model.

4. The method for characterizing small-scale fracture development zones based on a high-precision velocity model according to claim 3 is characterized in that: Correlation analysis and velocity correction of seismic layer velocity and layer velocity at well point are carried out, including: The cross-correlation analysis function is used to analyze the cross-correlation between the seismic layer velocity and the layer velocity at the well point, so as to obtain the difference between the seismic layer velocity and the layer velocity at the well point in different strata; Based on the difference between the seismic layer velocity and the well point layer velocity of different strata, a correction function is constructed by using the layer velocity value of the well and the seismic layer velocity value, so as to correct the seismic layer velocity value by using the correction function; The corrected seismic layer velocities are sequentially subjected to three-dimensional gridding processing to obtain a high-precision velocity model through thinning, interpolation and smoothing.

5. The method for characterizing small-scale fracture development zones based on a high-precision velocity model according to claim 1 is characterized in that: The process of converting the adaptive normal illumination data in the time domain into the adaptive normal illumination data in the depth domain using the high-precision velocity model is expressed as: F (h) = V vol ΔγC F (t), where C F (t) is the adaptive normal illumination data in the time domain, Δγ is the adaptive factor, C F (h) is the adaptive normal illumination data in the depth domain, and Vvol is the seismic layer velocity function in the high-precision velocity model.

6. The method for characterizing small-scale fracture development zones based on a high-precision velocity model according to claim 1, characterized in that: The initial fracture grid model and dynamic small-scale fracture grid model are obtained by using the orientation guidance field and fracture density optimized layer by layer based on the neural network unit, including: The crack density function and crack orientation function are constructed based on the crack circular discrete space function, the crack circular space radius and the vertical and horizontal coordinates; The fracture density function is normalized according to the maximum and minimum values ​​of the fracture density, the normalized fracture density and fracture orientation are subjected to high-density gridding, and the orientation guidance field is obtained; A neural network unit is established according to the azimuth guidance field function, and the azimuth guidance field in the neural network unit is corrected layer by layer using the azimuth guidance field functions of different wells, and an initial small-scale fracture grid model is obtained according to the azimuth guidance field and fracture density corrected layer by layer; According to the initial small-scale fracture grid model, model weight allocation, optimal weight value screening and high-density grid connection are carried out in sequence to output a dynamic small-scale fracture grid model suitable for the study area.

7. The method for characterizing small-scale fracture development zones based on a high-precision velocity model according to claim 1 or 6, characterized in that: The fracture development area is determined based on the density and orientation of small-scale fractures, including: The fracture density and fracture orientation are linearly regressed, and the scope and law of the small-scale fracture development zone are determined by combining the planar distribution characteristics of the fracture development zone in the dynamic small-scale fracture grid model.

8. A small-scale fracture development zone characterization system based on a high-precision velocity model, characterized in that: It includes: data acquisition module, data optimization module, speed model acquisition module, data conversion module, target model establishment deep learning module and regional representation module, among which, Data acquisition module, used to obtain pre-stack OVT domain seismic gather data and perform noise analysis; A data optimization module, used for AVO fitting compensation gather optimization processing, wherein the AVO fitting compensation gather optimization processing includes fitting a cubic polynomial, removing noise and target gather energy compensation; The velocity model acquisition module is used to correct the seismic layer velocity function of the target area using the logging data, and obtain the layer velocity information using the layer velocity function to establish a high-precision velocity model; A data conversion module, used for obtaining time-domain adaptive normal lighting data by using a multi-parameter optimized connectivity attribute body, and converting the time-domain adaptive normal lighting data into depth-domain adaptive normal lighting data by using a high-precision velocity model; The target model establishes a deep learning module, which is used to establish an orientation guidance field using fracture orientation and fracture density, and to correct the orientation guidance field in the neural network unit layer by layer using the orientation guidance fields of different wells. An initial small-scale fracture grid model is established using the corrected orientation guidance field and fracture density, and the model weight allocation, optimal weight value screening and high-density grid connection are sequentially performed on the initial small-scale fracture grid model to obtain a dynamic small-scale fracture grid model, which is used to characterize the fracture development area and spatial distribution characteristics. The regional characterization module is used to determine the fracture development area according to the small-scale fracture density and fracture orientation and based on the dynamic small-scale fracture grid model to characterize the small-scale fracture development area.

9. An electronic device, characterized in that: include: at least one processor, and a memory coupled to the at least one processor; The memory stores a computer program, and the computer program can be executed by the at least one processor to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed, the method according to any one of claims 1 to 7 can be implemented.

Images