A pre-stack fracture prediction method based on phased analysis and forward modeling

Abstract

The present invention relates to a pre-stack fracture prediction method based on phased analysis and forward modeling, belonging to the technical field of pre-stack fracture prediction. The method first classifies the completed drilled wells based on the fracture development situation, then classifies the seismic waveforms of the target layer using post-stack seismic data, establishes the corresponding relationship between the seismic waveforms and the completed drilled wells, then extracts the amplitude-azimuth curve of a single CMP point along the target layer, and calculates the amplitude-azimuth simulation curve under different fracture densities based on the completed drilled well data corresponding to the CMP point. Then, the R<supgt;2< / supgt; formula is used to calculate the correlation between the simulated and actual amplitude-azimuth curves, and the fracture density with the largest correlation is preferably selected as the fracture density of the current CMP point. Through the cyclic calculation of CMP points, the fracture density of the target layer in the study area is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas field exploration, and particularly relates to a pre-stack fracture prediction method based on phase-controlled analysis and forward modeling. Background Technique

[0002] In seismic exploration in the field of oil and gas field exploration, fractures are very important information and play a crucial role in oil and gas enrichment and reservoir formation. However, due to the multi-stage and diverse distribution laws, genetic types, filling conditions, etc. of fractures during the tectonic evolution process, fractures of different periods overlap, restrict, and transform each other, increasing the complexity of the spatial distribution of fractures, that is, anisotropy. Therefore, in reservoir prediction dominated by fractures, it is often necessary to identify and characterize the strike, density, and distribution range of underground fractures through seismic data, and this process is called fracture prediction.

[0003] Currently, the commonly used pre-stack fracture prediction methods at home and abroad mainly include shear wave splitting, converted wave detection, and P-wave anisotropy detection techniques, etc. Among them, the P-wave anisotropy detection technique is the most widely used for high-angle micro-fractures.

[0004] I. Principle of Shear Wave Splitting Method

[0005] The shear wave splitting phenomenon is the most direct response of seismic waves to fractures and is sensitive to the existence of fractures, which determines that using shear waves to detect fractures is relatively accurate and effective. When the propagation direction of the shear wave is oblique to the fractures, the shear wave can be split into a fast shear wave (called S1 wave) polarized along the fracture development direction and a slow shear wave (called S2 wave) polarized along the orthogonal direction. In order to observe the shear wave splitting phenomenon, shear wave data must be obtained. Currently, it is mainly obtained through shear wave exploration, shear wave four-component exploration, and converted wave exploration. The first two methods require too high exploration costs, and the latter method requires collecting shear wave information with a high signal-to-noise ratio.

[0006] II. Principle of P-Wave Anisotropy Detection Technique

[0007] The azimuthal anisotropy of seismic waves refers to the fact that the characteristics of the seismic wave field will change with the change of the propagation direction or the observation direction. According to the research of wave theory, seismic waves will produce certain anisotropic characteristics when passing through the fracture development zone, including amplitude variation with azimuth (AVAZ), apparent velocity variation with azimuth (VVAZ), wavelet frequency variation with azimuth (FVAZ), frequency attenuation variation with azimuth (QVAZ), etc. These anisotropic characteristics make it possible to detect the azimuth and development density of fracture development by detecting these changes or anomalies. The main theory of this method is the Rüger formula, but its assumption condition is the EDA medium, that is, parallel arranged vertical fractures. Therefore, further in-depth research is still needed on the prediction theory of medium and low-angle fractures and the issue of the selection and rejection of the last two terms of the Rüger formula.

[0008] III. Defects of Existing Methods

[0009] 1. The existing shear wave splitting method is mainly obtained through converted waves. However, the horizontal component data collected often has lower energy and signal-to-noise ratio than the vertical component. Therefore, for the exploration of fractured formations with azimuthal anisotropy, not only the parameters of seismic acquisition need to be carefully designed (such as the directions of main survey lines and connecting survey lines, the aspect ratio of the observation system, the azimuthal coverage times, etc.), but also higher requirements are imposed on the accuracy of acquisition equipment and the amplitude-preserving processing and vector processing of multi-component seismic data. This restricts the development of converted wave fracture detection.

[0010] 2. The main theory of the existing longitudinal wave anisotropy detection technology is the Rüger formula. The assumption conditions of this formula are for EDA media, while the actual formation often does not have vertically fractured cracks arranged in parallel. This causes certain deficiencies in the selection of fracture prediction models, which will reduce the accuracy of fracture prediction to a certain extent.

[0011] Therefore, in view of the common phenomenon of strong pre-stack fracture anisotropy, that is, the phenomenon that pre-stack fractures overlap, restrict, and transform each other in space, the present invention invents a pre-stack fracture prediction method based on phased analysis and forward modeling. Under the guidance of phased analysis, forward modeling is used to analyze the correlation of AVAZ curves for fracture prediction. Summary of the Invention

[0012] The present invention relates to a pre-stack fracture prediction method based on phased analysis and forward modeling, belonging to the technical field of pre-stack fracture prediction. The method first classifies the wells in the study area based on the fracture development situation of the completed wells drilled through the target layer, and then classifies the seismic waveforms of the target layer using post-stack seismic data. The number of seismic waveform classifications is the same as the number of completed well classifications, and the corresponding relationship between seismic waveforms and completed wells is established. Then, AVAZ gather analysis of a single CMP point is carried out. First, the amplitude-azimuth angle curve of this CMP point is extracted along the target layer, and then based on the coordinates of this CMP point and the seismic waveform classification results, the completed well corresponding to this CMP point is identified, and the amplitude-azimuth angle simulation curves under different fracture densities are calculated using the data of this completed well. Then, the correlation between the simulation and the actual amplitude-azimuth angle curves is calculated using the R 2 formula, and the fracture density with the maximum correlation is selected as the fracture density of the current CMP point. Through the cyclic calculation of CMP points, the fracture density of the target layer in the study area is obtained.

[0013] The specific steps of the present invention include:

[0014] (1) Input the logging data of all completed wells (drilling through the target layer) in the study area. According to the different fracture development situations represented by the completed wells, the completed wells are divided into N categories, and the number of categories of completed wells is not less than 3 categories;

[0015] (2) Input the three-dimensional post-stack data volume of the study area. For the target interval, use the seismic waveform classification method to divide this interval into N types of seismic waveforms on the plane, and establish the corresponding relationship between the seismic waveforms and the drilled wells, that is, the i-th type of waveform corresponds to the i-th type of drilled well;

[0016] (3) For a single CMP point within the target interval:

[0017] (3-1) Extract the AVAZ gather of the current CMP point along the target interval from the pre-stack seismic data of the study area, and extract the AVAZ curve, that is, the actual AVAZ curve, with the ordinate being the amplitude and the abscissa being the azimuth;

[0018] (3-2) Combine steps (1), (2) and step (3-1) to identify the drilled well corresponding to this CMP point based on the coordinates of this CMP point and the seismic waveform classification results;

[0019] (3-3) Based on the logging data of the drilled well obtained in step (3-2), calculate the AVAZ curves under different fracture densities, that is, the AVAZ simulation curves;

[0020] (3-4) Use the R 2 formula to calculate the correlation between the simulated AVAZ curve in step (3-3) and the actual AVAZ curve in step (3-1) in turn, and select the fracture density with the maximum correlation as the fracture prediction density of the current CMP point;

[0021] (4) Loop through all CMP points in the current interval to obtain the fracture density prediction data.

[0022] A pre-stack fracture prediction method based on phased analysis and forward modeling has the following characteristics, mainly manifested as:

[0023] (1) The present invention uses the idea of phased control to establish the connection between a certain CMP point in the pre-stack gather and the typical drilled well through the fracture development situation of the drilled well and the seismic waveforms of the target interval; by calculating the correlation between the forward curve of the drilled well data and the actual formation curve, the fracture development density of the actual formation is obtained, which has strong innovation.

[0024] (2) The core basis of the present invention lies in how to establish the connection between the completed well and the CMP point of the actual target layer and whether the forward modeling result of AVAZ can be used to indicate the fracture development density. Through years of research, the inventor found that different fracture densities and different seismic waveforms are developed in the completed wells, and different fracture densities and seismic waveforms are also developed in the actual target layer. By calculating the similarity of the waveforms, the connection between the CMP point and the completed well can be established; by calculating the correlation between the forward modeling of AVAZ with different fracture densities and the AVAZ curve of the pre-stack seismic gather, pre-stack fracture prediction can be further carried out; and there is theoretical support for the method: ① The seismic waveform of the completed well is a part of the actual seismic waveform, and a one-to-one correspondence relationship between the waveform of the completed well and the actual seismic waveform can be found; ② Using the logging data of the completed well for forward modeling of AVAZ can approximately represent the AVAZ curve of the pre-stack seismic gather with the same fracture density as that of the well. Therefore, the present invention has a good theoretical basis for the method. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is the technical flow chart of the present invention;

[0026] Figure 2 is the pre-stack fracture prediction plan view obtained by using the conventional method;

[0027] Figure 3 is the pre-stack fracture prediction plan view obtained by using the method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0028] EXAMPLE

[0029] A pre-stack fracture prediction method based on phased analysis and forward modeling, the steps include:

[0030] Step 1: Input the logging data of all the completed wells in the study area. According to the different fracture density characteristics represented by the completed wells, the completed wells are divided into N categories, W N = {W 1 , W 2 , W 3 , …, W N}, and the number of categories of the completed wells is not less than 3.

[0031] Step 2: Input the stacked seismic data of the study area, denoted as Sh(x, y, t), where x represents the lateral survey line number, y represents the longitudinal survey line number, h represents the offset, t represents the time, and denote the time of the target layer as from t0 to t1; for the time period from t0 to t1 of the target layer section, perform seismic waveform classification analysis on the post-stack seismic data of this layer section. The number of waveform classifications is set to N categories, which respectively represent different seismic facies, and denote the corresponding waveform classification array as B(x, y) = {B 1 , B 2 , B 3 , …, B N}, the earthquake waveform classification method uses the k-means clustering method and establishes the corresponding relationship between the earthquake waveform and the completed drilled well, that is, the i-th type of waveform corresponds to the i-th type of completed drilled well.

[0032] Step 3: For the pre-stack seismic gather of a certain CMP point within the current small interval, the planar coordinates of this CMP point are x and y, and its amplitude energy is denoted as , h is the offset, is the azimuth angle, t is the time,

[0033] Step 3-1: Extract the AVAZ gather of the CMP point along the target layer from the pre-stack seismic data in the study area, and extract the amplitude curves at different azimuth angles along t0 The relationship between the amplitude and the azimuth angle is denoted as , is the azimuth angle,

[0034] Step 3-2: Identify the completed drilled well corresponding to this CMP point based on the planar coordinates (x, y) of this CMP point and the earthquake waveform classification results. Assume it is Well W 1 Well.

[0035] Step 3-3: Use the logging data of Well W in step (3-2) to perform AVAZ forward modeling. The fracture densities are set to 0%, 2%, 4%, …, 20% in sequence. The forward modeling formula is calculated using the formula for the reflection coefficient R of P-waves in azimuthally anisotropic media derived by Rüger (1997) as a function of the incident angle and azimuth angle, and obtain the simulated array of amplitude-azimuth angles for different fracture densities of Well W 1 Well, 1 where , is the azimuth angle, i is different fracture densities (0%, 2%, 4%, …, 20%),

[0036] ;

[0037] In Formula 1, is the P-wave reflection coefficient, is the incident angle, p is the density, and are the wave impedance of the P-wave in the vertical direction and the modulus of the S-wave; and are respectively the average values of the vertical P-wave and S-wave velocities on both sides of the interface,

[0038] Step 3-4: Use the R 2 formula to calculate the correlation analysis between the fracture density (i = 0%, i = 2%, i = 4%, …, i = 20%) and the AVAZ curve obtained in step (3-1) When R 2When it is the largest, the i at this time is recorded as the fracture density of this CMP point, denoted as lf(x, y);

[0039] ;

[0040] In formula 2, i is the fracture density, is the azimuth angle, is the amplitude value when the forward modeling density is i and the azimuth angle is 0, is the amplitude value when the azimuth angle is 0 along the AVAZ curve extracted at t0,

[0041] Step 4: Replace the CMP point number, repeat step (3), loop through all CMP points within the current interval, and obtain the fracture prediction plan view lf(line, xline) after fracture prediction, where line is the line number and xline is the trace number.

[0042] Embodiment

[0043] Figure 2 and Figure 3 are respectively embodiments of pre-stack fracture prediction for the target interval in a certain research area. Figure 2 is the plan view of the conventional pre-stack fracture prediction result in the research area. Figure 3 By using the fracture prediction plan view obtained by the present invention, it can be seen that the fracture prediction effect by using the present invention is better.

[0044] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A pre-stack fracture prediction method based on phase control analysis and forward modeling, comprising the following steps: Step 1: Input the logging data of all completed wells in the study area, and divide the completed wells into N categories, W according to the different fracture density characteristics represented by the completed wells. N ={W1,W2,W3,…,W N }, the number of completed wells is not less than 3; Step 2: Input the stacked seismic data of the study area, recorded as Sh(x,y,t), where x represents the horizontal survey line number, y represents the vertical survey line number, h represents the offset distance, and t represents the time. The target layer time is recorded as t0 to t1; For the target layer from time t0 to t1, seismic waveform classification analysis is performed on the layer of the post-stack seismic data. The number of waveform classifications is set to N, representing different seismic phases. The corresponding waveform classification array is B(x, y) = {B1, B2, B3, …, B N }, the seismic waveform classification method adopts the k-means clustering method, and establishes the corresponding relationship between the seismic waveform and the completed well, that is, the i-th type of waveform corresponds to the i-th type of completed well; Step 3: For the seismic pre-stack gather of a CMP point in the current small layer segment, the plane coordinates of the CMP point are x and y, and its amplitude energy is recorded as Aq(h, ,t), is the offset distance, is the azimuth, t is the time; Step 3-1: Extract the AVAZ gathers of CMP points along the target layer from the pre-stack seismic data of the study area, and extract the AVAZ gathers of different azimuths along t0. The amplitude curve of the vibration is recorded as the relationship between the amplitude and the azimuth angle. , is the azimuth; Step 3-2: Identify the completed well corresponding to the CMP point based on the plane coordinates (x, y) of the CMP point and the seismic waveform classification results; Step 3-3: Use logging data to perform AVAZ forward modeling. The fracture density is set to 0%, 2%, 4%, 6%, ..., 20% in sequence. The forward modeling formula is calculated using formula (1) to obtain the amplitude-azimuth simulation array of different fracture densities. , is the azimuth angle, i is the different crack densities; ； In formula (1), is the P-wave reflection coefficient, is the incident angle, is the density, and is the wave impedance of the P wave and the modulus of the S wave in the vertical direction; and are the average values ​​of the vertical P-wave and S-wave velocities on both sides of the interface, , and are the anisotropic parameters of the strata above and below the interface , and The difference, Step 3-4: Using R 2 The formula calculates the crack density and the AVAZ curve obtained in step (3-1) For correlation analysis, when R 2 When it is at its maximum, i at this time is recorded as the crack density of the CMP point, denoted as lf(x, y); ； In formula (2), i is the crack density, is the azimuth, for The mean of for The mean of Step 4: Change the CMP point number and repeat step (3) to loop through all CMP points in the current layer segment to obtain the fracture prediction plane diagram lf(line, xline) after fracture prediction, where line is the line number and xline is the track number.

Images