Virtual well construction method and device based on seismic data mining

Abstract

The invention relates to the technical field of seismic data interpretation, and particularly discloses a virtual well construction method and device based on seismic data mining, and the method comprises the steps: constructing a seismic reflection mapping network through employing a reflection coefficient of a known well and a seismic channel beside the well; calculating an initial virtual well reflection coefficient by using the seismic reflection mapping network; based on the initial virtual well reflection coefficient, solving a target virtual well reflection coefficient through a space matching operator; and based on the target virtual well reflection coefficient, calculating through an impedance reflection coefficient formula to obtain target virtual well impedance. According to the virtual well construction method based on seismic data mining, an initial virtual well reflection coefficient is solved by using a seismic reflection mapping network, a target virtual well reflection coefficient is solved through a space matching operator, and finally, target virtual well impedance is calculated according to an impedance reflection coefficient formula. The problem that the seismic inversion precision is finally affected due to low interpolation modeling precision caused by non-uniformly distributed logging data is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of seismic data interpretation, and particularly relates to a virtual well construction method and device based on seismic data mining. BACKGROUND

[0002] Traditional ground seismic inversion (for example, wave impedance inversion) needs to be interpolated by taking well logging data and seismic data as longitudinal and horizontal control to obtain an initial model. However, in the actual data inversion process, the actual exploration and development well point distribution of the whole work area is often uneven. The unevenly distributed well logging data can result in low precision of interpolation modeling, and finally affect the seismic inversion precision.

[0003] In order to solve the contradiction between the demand of seismic inversion for well logging data in the modeling process and the uneven distribution of actual exploration and development well points, the present application provides a virtual well construction method based on seismic data mining. The initial virtual well reflection coefficient is obtained by using a seismic reflection mapping network, the target virtual well reflection coefficient is obtained by using a spatial matching operator, and finally the target virtual well impedance is calculated according to the impedance reflection coefficient formula. Further, the virtual well data is increased to achieve the goal of data enhancement, improve the precision of interpolation modeling, and improve the seismic inversion precision. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a virtual well construction method and device based on seismic data mining. The method uses a seismic reflection mapping network to obtain an initial virtual well reflection coefficient, uses a spatial matching operator to obtain a target virtual well reflection coefficient, and finally calculates a target virtual well impedance according to an impedance reflection coefficient formula, thereby solving the problem that unevenly distributed well logging data can result in low precision of interpolation modeling and finally affect the seismic inversion precision.

[0005] In order to achieve the above-mentioned purpose, the first aspect of the present application provides a virtual well construction method based on seismic data mining, comprising:

[0006] A seismic reflection mapping network is constructed by using the reflection coefficient of a known well and a seismic trace beside the well;

[0007] An initial virtual well reflection coefficient is obtained by using the seismic reflection mapping network;

[0008] A target virtual well reflection coefficient is obtained by using a spatial matching operator based on the initial virtual well reflection coefficient;

[0009] A target virtual well impedance is calculated by using an impedance reflection coefficient formula based on the target virtual well reflection coefficient.

[0010] The second aspect of the present application provides a virtual well construction device based on seismic data mining, comprising:

[0011] a network constructing module, configured to construct a seismic reflection mapping network by using the reflection coefficients of known wells and seismic traces beside the wells;

[0012] an initial coefficient obtaining module, configured to obtain initial virtual well reflection coefficients by using the seismic reflection mapping network;

[0013] a target coefficient obtaining module, configured to obtain target virtual well reflection coefficients by a spatial matching operator based on the initial virtual well reflection coefficients;

[0014] an impedance calculating module, configured to calculate target virtual well impedance by an impedance reflection coefficient formula based on the target virtual well reflection coefficients.

[0015] A third aspect of the present application provides an electronic device, comprising:

[0016] a memory, which stores executable instructions;

[0017] a processor, which runs the executable instructions in the memory to realize the virtual well construction method based on seismic data mining in the first aspect.

[0018] A fourth aspect of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the virtual well construction method based on seismic data mining in the first aspect.

[0019] The present application has the following beneficial effects:

[0020] (1) The virtual well construction method based on seismic data mining provided by the present application obtains initial virtual well reflection coefficients by using a seismic reflection mapping network, obtains target virtual well reflection coefficients by a spatial matching operator, and finally calculates target virtual well impedance according to an impedance reflection coefficient formula, thereby solving the problem that unevenly distributed logging data can result in low precision of interpolation modeling and finally affect the precision of seismic inversion.

[0021] (2) The virtual well construction method based on seismic data mining provided by the present application has the characteristics of easy implementation, strong robustness and wide adaptability compared with traditional virtual well generation methods, and is very suitable for logging data preparation for seismic inversion.

[0022] Other features and advantages of the present application will be described in detail in the following specific embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0023] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which:

[0024] Figure 1 This is a flowchart illustrating the virtual well construction method based on seismic data mining proposed in this invention.

[0025] Figure 2 This is a schematic diagram illustrating the process of obtaining the initial virtual well reflection coefficient using a seismic reflection mapping network in a specific embodiment of the virtual well construction method based on seismic data mining proposed in this invention.

[0026] Figure 3 This is a schematic diagram illustrating the determination of the reflection coefficient of a target virtual well using a spatial matching operator in a specific implementation of the virtual well construction method based on seismic data mining proposed in this invention.

[0027] Figure 4 This is a schematic diagram of the final target virtual well impedance obtained in a specific embodiment of the virtual well construction method based on seismic data mining proposed in this invention. Detailed Implementation

[0028] Preferred embodiments of the invention will now be described in more detail. While preferred embodiments of the invention are described below, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0029] This invention provides a method for constructing virtual wells based on seismic data mining, such as... Figure 1 As shown, it includes:

[0030] A seismic reflection mapping network is constructed using the known reflection coefficients of wells and seismic traces near the wells;

[0031] The initial virtual well reflection coefficients are obtained using a seismic reflection mapping network;

[0032] Based on the initial virtual well reflection coefficient, the target virtual well reflection coefficient is obtained using a spatial matching operator;

[0033] The impedance of the target virtual well is calculated using the impedance reflection coefficient formula based on the reflection coefficient of the target virtual well.

[0034] In this invention, the initial virtual well reflection coefficient is obtained using a seismic reflection mapping network, the target virtual well reflection coefficient is obtained using a spatial matching operator, and finally the target virtual well impedance is calculated according to the impedance reflection coefficient formula. This solves the problem that unevenly distributed well logging data can lead to low accuracy in interpolation modeling, which ultimately affects the accuracy of seismic inversion.

[0035] According to the present invention, constructing a seismic reflection mapping network using known well reflection coefficients and well-side seismic traces includes:

[0036] Wavelet is extracted using the known reflection coefficient of the well and the seismic traces around the well, according to the convolution model;

[0037] Multiple groups of reflection coefficients are generated by randomly shuffling the order of the reflection coefficients.

[0038] Multiple sets of reflection coefficients are convolved with wavelet to obtain multiple sets of seismic traces;

[0039] Using well-side seismic traces and multiple seismic trace groups as inputs, and reflection coefficients and multiple reflection coefficient groups as outputs, a deep network is trained to obtain a seismic reflection mapping network.

[0040] According to the present invention, obtaining the initial virtual well reflection coefficient using a seismic reflection mapping network includes:

[0041] The seismic data at the target point location is input into the seismic reflection mapping network to obtain the initial virtual well reflection coefficient at the target point location.

[0042] According to the present invention, obtaining the target virtual well reflection coefficient based on the initial virtual well reflection coefficient using a spatial matching operator includes:

[0043] Substitute the initial virtual well reflection coefficient into the spatial matching operator expression, and use the gradient descent method to find the minimum of the spatial matching operator to obtain the target virtual well reflection coefficient.

[0044] Preferably, the expression for the spatial matching operator is:

[0045] E = ||I w2 -flatten(I w1 )||2+||R w2 -R model (w2)||0+||curv(I w2 )-curv(I w1 )||2;

[0046] Among them, I w2 For the impedance of the target virtual well, flatten(I) w1 R is a flattening function for known well impedance. w2 R represents the reflection coefficient of the target virtual well. model (w2) is the initial virtual well reflection coefficient, curv(I) w2 ) represents the curvature of the target virtual well impedance, curv(I) w1 ) represents the curvature of the known well impedance.

[0047] According to the present invention, the flattening of the known well impedance flattening function includes longitudinal flattening and lateral displacement;

[0048] The longitudinal flattening references the geological interpretation strata, while the lateral displacement represents the target virtual well location.

[0049] Preferably, the formula for the impedance reflection coefficient is:

[0050] I = e 2 ∫Rdt;

[0051] Where R is the reflection coefficient.

[0052] Compared with traditional virtual well generation methods, this invention is easier to implement, more robust, and more adaptable, making it highly suitable for preparing well logging data for seismic inversion.

[0053] The present invention will be described in more detail below through embodiments.

[0054] Example 1:

[0055] like Figures 2-4 As shown in the figure, this embodiment proposes a virtual well construction method based on seismic data mining. This method is based on the spatial matching operator, and the calculation formula of the spatial matching operator is as follows:

[0056] E = ||I w2 -flatten(I w1 )||2+||R w2 -R model (w2)||0+||curv(I w2 )-curv(I w1 )||2;

[0057] Among them, I w2 For the impedance of the target virtual well, flatten(I) w1 R is a flattening function for known well impedance, representing the flattening process applied to known well impedances. Vertical flattening is referenced to the geological interpretation stratigraphic level, and horizontal displacement is the target virtual well location. w2 R represents the reflection coefficient of the target virtual well. model (w2) is the initial virtual well reflection coefficient, curv(I) w2 ) represents the curvature of the target virtual well impedance, curv(I) w1 () represents the curvature of the known well impedance;

[0058] The first term of the spatial matching operator ensures that the impedance of the target virtual well can be influenced by the impedance of nearby known wells, so that the generated virtual well conforms to the underground geological laws; the third term, for the calculation of well impedance, matches the trend of the well curve concavity and convexity, thereby achieving the purpose of removing curve outliers.

[0059] The specific implementation steps of this method are as follows:

[0060] (1) Calculate the reflection coefficient R of the known well W1. w1 ;

[0061] (2) Using the known reflection coefficient R of well W1 w1The wavelet w was extracted from the well-side seismic trace S1 according to the convolution model;

[0062] (3) The reflection coefficient R of the known well W1 calculated in step (1) w1 Randomly shuffle the order to generate N additional reflection coefficient groups R*;

[0063] (4) The reflection coefficient group R generated in step (3) * Convolved with the wavelet w extracted in step (2), N additional seismic trace groups S are obtained. * ;

[0064] (5) Utilizing the known well-side trace S1 of well W1 and the added seismic trace group S * As input, the reflection coefficient R of well W1 is known. w1 And the increased reflection coefficient group R * As output, a deep network is trained; this yields a seismic reflection mapping network.

[0065] (6) Using the seismic data at the target point W2 as input, input the seismic reflection mapping network obtained in step (5) to obtain the initial virtual well reflection coefficient R at the target point W2. model (w2);

[0066] (7) The initial virtual well reflection coefficient R obtained in step (6) model Substituting (w2) into the spatial matching operator calculation formula, the gradient descent method is used to find the minimum of the spatial matching operator, thus obtaining the reflection coefficient R of the target virtual well W2. w2 ;

[0067] According to the formula for impedance reflection coefficient, I = e 2∫Rdt The well impedance I of the target virtual W2 was calculated. w2 .

[0068] In this embodiment, the above-mentioned virtual well construction method based on seismic data mining is verified as follows:

[0069] 1) Using the seismic data at target point W2 as input, input it into the seismic reflection mapping network to obtain the initial virtual well reflection coefficient R at the target point location. model (w2). For example... Figure 2 As shown;

[0070] 2) Given the well impedance I w1 And 1) the initial virtual well reflection coefficient R obtained. model Substituting (w2) into formula 1), the gradient descent method is used to find the minimum of the spatial matching operator, and the reflection coefficient R of the target virtual well W2 is obtained. w2 ;like Figure 3 As shown;

[0071] 3) According to the impedance reflection coefficient formula I = e 2∫Rdt The target virtual well impedance I was calculated. w2 ;like Figure 4 As shown.

[0072] Example 2:

[0073] This embodiment provides a virtual well construction method based on seismic data mining, such as... Figure 1 As shown, it includes:

[0074] A seismic reflection mapping network is constructed using the known reflection coefficients of wells and seismic traces near the wells;

[0075] The initial virtual well reflection coefficients are obtained using a seismic reflection mapping network;

[0076] Based on the initial virtual well reflection coefficient, the target virtual well reflection coefficient is obtained using a spatial matching operator;

[0077] The impedance of the target virtual well is calculated based on the reflection coefficient of the target virtual well using the impedance reflection coefficient formula.

[0078] Constructing a seismic reflection mapping network using known well reflection coefficients and well-side seismic traces includes:

[0079] Wavelet is extracted using the known reflection coefficient of the well and the seismic traces around the well, according to the convolution model;

[0080] Multiple groups of reflection coefficients are generated by randomly shuffling the order of the reflection coefficients.

[0081] Multiple sets of reflection coefficients are convolved with wavelet to obtain multiple sets of seismic traces;

[0082] Using well-side seismic traces and multiple seismic trace groups as inputs, and reflection coefficients and multiple reflection coefficient groups as outputs, a deep network is trained to obtain a seismic reflection mapping network.

[0083] The initial virtual well reflection coefficients are obtained using a seismic reflection mapping network, including:

[0084] The seismic data at the target point location is input into the seismic reflection mapping network to obtain the initial virtual well reflection coefficient at the target point location;

[0085] Based on the initial virtual well reflection coefficient, the target virtual well reflection coefficient is obtained using the spatial matching operator, including:

[0086] Substitute the initial virtual well reflection coefficient into the spatial matching operator expression, and use the gradient descent method to find the minimum of the spatial matching operator to obtain the target virtual well reflection coefficient;

[0087] The expression for the spatial matching operator is:

[0088] E = ||I w2 -flatten(I w1 )||2+||R w2 -R model (w2)||0+||curv(I w2 )-curv(I w1 )||2;

[0089] Among them, I w2 For the impedance of the target virtual well, flatten(I) w1 R is a flattening function for known well impedance. w2 R represents the reflection coefficient of the target virtual well. model (w2) is the initial virtual well reflection coefficient, curv(I) w2 ) represents the curvature of the target virtual well impedance, curv(I) w1 () represents the curvature of the known well impedance;

[0090] The flattening of the well impedance flattening function is known to include longitudinal flattening and lateral displacement;

[0091] The longitudinal horizontal alignment references the geological interpretation stratigraphic level, while the lateral displacement represents the target virtual well location.

[0092] The formula for impedance reflection coefficient is:

[0093] I = e 2∫Rdt ;

[0094] Where R is the reflection coefficient.

[0095] Example 3:

[0096] This embodiment provides a virtual well construction device based on seismic data mining, including:

[0097] A network building module is used to construct a seismic reflection mapping network using known well reflection coefficients and well-side seismic traces;

[0098] The initial coefficients calculation module is used to calculate the initial virtual well reflection coefficients using a seismic reflection mapping network;

[0099] The target coefficient calculation module is used to calculate the target virtual well reflection coefficient based on the initial virtual well reflection coefficient using a spatial matching operator.

[0100] The impedance calculation module is used to calculate the impedance of the target virtual well based on the reflection coefficient of the target virtual well using the impedance reflection coefficient formula.

[0101] Constructing a seismic reflection mapping network using known well reflection coefficients and well-side seismic traces includes:

[0102] Wavelet is extracted using the known reflection coefficient of the well and the seismic traces around the well, according to the convolution model;

[0103] Multiple groups of reflection coefficients are generated by randomly shuffling the order of the reflection coefficients.

[0104] Multiple sets of reflection coefficients are convolved with wavelet to obtain multiple sets of seismic traces;

[0105] Using well-side seismic traces and multiple seismic trace groups as inputs, and reflection coefficients and multiple reflection coefficient groups as outputs, a deep network is trained to obtain a seismic reflection mapping network.

[0106] The initial virtual well reflection coefficients are obtained using a seismic reflection mapping network, including:

[0107] The seismic data at the target point location is input into the seismic reflection mapping network to obtain the initial virtual well reflection coefficient at the target point location;

[0108] Based on the initial virtual well reflection coefficient, the target virtual well reflection coefficient is obtained using the spatial matching operator, including:

[0109] Substitute the initial virtual well reflection coefficient into the spatial matching operator expression, and use the gradient descent method to find the minimum of the spatial matching operator to obtain the target virtual well reflection coefficient;

[0110] The expression for the spatial matching operator is:

[0111] E = ||I w2 -flatten(I w1 )||2+||R w2 -R model (w2)||0+||curv(I w2 )-curv(I w1 )||2;

[0112] Among them, I w2 For the impedance of the target virtual well, flatten(I) w1 R is a flattening function for known well impedance. w2 R represents the reflection coefficient of the target virtual well. model (w2) is the initial virtual well reflection coefficient, curv(I) w2 ) represents the curvature of the target virtual well impedance, curv(I) w1 () represents the curvature of the known well impedance;

[0113] The flattening of the well impedance flattening function is known to include longitudinal flattening and lateral displacement;

[0114] The longitudinal horizontal alignment references the geological interpretation stratigraphic level, while the lateral displacement represents the target virtual well location.

[0115] The formula for impedance reflection coefficient is:

[0116] I = e 2∫Rdt ;

[0117] Where R is the reflection coefficient.

[0118] Example 4:

[0119] This invention provides an electronic device including a memory and a processor.

[0120] Memory, which stores executable instructions;

[0121] The processor executes executable instructions in memory to implement a virtual well construction method based on seismic data mining.

[0122] This memory is used to store non-transitory computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0123] The processor may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. In one embodiment of the invention, the processor is used to execute computer-readable instructions stored in the memory.

[0124] Those skilled in the art should understand that, in order to solve the technical problem of how to achieve a good user experience, this embodiment may also include well-known structures such as communication buses and interfaces, and these well-known structures should also be included within the protection scope of this invention.

[0125] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0126] Example 5:

[0127] This invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements a virtual well construction method based on seismic data mining.

[0128] A computer-readable storage medium according to embodiments of the present invention stores non-transitory computer-readable instructions. When these non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the methods described in the foregoing embodiments of the present invention are performed.

[0129] The aforementioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or portable hard drive), media with built-in rewritable non-volatile memory (e.g., memory card), and media with built-in ROM (e.g., ROM cartridge).

[0130] The virtual well construction method based on seismic data mining proposed in the embodiments of the present invention uses a seismic reflection mapping network to obtain the initial virtual well reflection coefficient, uses a spatial matching operator to obtain the target virtual well reflection coefficient, and finally calculates the target virtual well impedance according to the impedance reflection coefficient formula. This solves the problem that unevenly distributed well logging data will lead to low accuracy of interpolation modeling, which will ultimately affect the accuracy of seismic inversion.

[0131] The various embodiments of the present invention have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A method for constructing a virtual well based on seismic data mining, characterized in that, The method comprises the following steps: constructing a seismic reflection mapping network by using the reflection coefficients of the known well and the seismic trace beside the well; obtaining initial virtual well reflection coefficients by using the seismic reflection mapping network; obtaining target virtual well reflection coefficients by using a spatial matching operator based on the initial virtual well reflection coefficients; calculating target virtual well impedance by using an impedance reflection coefficient formula based on the target virtual well reflection coefficients.

2. The method of claim 1, wherein, The method of constructing a seismic reflection mapping network by using the reflection coefficients of the known well and the seismic trace beside the well comprises the following steps: extracting a wavelet according to a convolution model by using the reflection coefficients of the known well and the seismic trace beside the well; generating a plurality of reflection coefficient groups by randomly shuffling the reflection coefficients; convolving the plurality of reflection coefficient groups with the wavelet to obtain a plurality of seismic trace groups; training a deep network by using the seismic trace beside the well and the plurality of seismic trace groups as input and the reflection coefficients and the plurality of reflection coefficient groups as output to obtain a seismic reflection mapping network.

3. The method of claim 2, wherein, The method of obtaining initial virtual well reflection coefficients by using the seismic reflection mapping network comprises the following steps: inputting seismic data of a target point position into the seismic reflection mapping network to obtain initial virtual well reflection coefficients of the target point position.

4. The method of claim 3, wherein, The method of obtaining target virtual well reflection coefficients by using a spatial matching operator based on the initial virtual well reflection coefficients comprises the following steps: obtaining the target virtual well reflection coefficients by substituting the initial virtual well reflection coefficients into the expression of the spatial matching operator and using a gradient descent method to minimize the spatial matching operator.

5. The method of claim 4, wherein, The expression of the spatial matching operator is: E = || I w2 -flatten(I w1 )||2 + || R w2 -R model (w2)||0 + || curv(I w2 )-curv(I w1 )||2; where I w2 is the impedance of the target virtual well, flatten(I w1 ) is a flattening function on the known well impedance, R w2 is the reflection coefficient of the target virtual well, R model (w2) is the initial virtual well reflection coefficient, curv(I w2 ) is the curvature of the target virtual well impedance, and curv(I w1 ) is the curvature of the known well impedance.

6. The method of claim 5, wherein, The flattening of the impedance of the known well comprises longitudinal flattening and lateral displacement. The longitudinal flattening is referred to as a geologically interpreted horizon, and the lateral displacement is the position of the target virtual well.

7. The method of claim 6, wherein, The impedance reflection coefficient formula is: I = e 2∫Rdt ; wherein R is the reflection coefficient.

8. A virtual well construction device based on seismic data mining, characterized by, The method comprises the following steps: a network construction module for constructing a seismic reflection mapping network by using the reflection coefficients of the known well and the seismic trace beside the well; an initial coefficient obtaining module for obtaining initial virtual well reflection coefficients by using the seismic reflection mapping network; a target coefficient obtaining module for obtaining target virtual well reflection coefficients by using a spatial matching operator based on the initial virtual well reflection coefficients; an impedance calculation module for calculating target virtual well impedance by using an impedance reflection coefficient formula based on the target virtual well reflection coefficients.

9. An electronic device, comprising: The electronic device comprises: a memory storing executable instructions; a processor running the executable instructions in the memory to implement the method of constructing a virtual well based on seismic data mining according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program which, when executed by a processor, implements the method of constructing a virtual well based on seismic data mining according to any one of claims 1-7.

Images