Multi-means integrated permeability quantitative prediction method, device, equipment and medium

Abstract

The invention belongs to the field of oil and gas geophysical exploration, and provides a multi-means integrated permeability quantitative prediction method, device and equipment and a medium. The prediction method comprises the following steps: acquiring seismic data, horizon data and logging data, and converting the seismic data, the horizon data and the logging data in the same domain; based on the seismic data, the horizon data and the logging data, respectively obtaining an interpolation permeability perminter, a rock physical permeability permrp and a machine learning permeability permml; and calculating the final multi-means integrated permeability permfinal on the basis of the obtained interpolation permeability perminter, the rock physical permeability permrp and the machine learning permeability permml. According to the method, earthquake, horizon and logging multivariate information is comprehensively utilized, three permeability prediction results are obtained respectively, the optimal weights of the three permeability prediction results are obtained based on the ant colony algorithm, and the final permeability quantitative prediction result is obtained through integration.

Description

Technical Field

[0001] The present invention belongs to the field of oil and gas geophysical exploration, and specifically relates to a method for quantitatively predicting permeability by integrating multiple means, a prediction device, an electronic device, and a storage medium. Background Art

[0002] Permeability is one of the important parameters characterizing reservoirs, and its accurate prediction can effectively improve the success rate of exploration and development. Seismic permeability prediction plays an important role in oil and gas reservoir development and reservoir evaluation. Using seismic data for reservoir permeability prediction can provide effective seepage field information for reservoir prediction, which is of great significance for reservoir evaluation and oil and gas reservoir development. Permeability is an important parameter for characterizing reservoir characteristics and the fluidity of underground fluids, and plays an important role in the design of development plans and the study of underground fluid characteristics. Therefore, it is of great significance to predict underground permeability parameters. However, the permeability response characteristics in conventional seismic data are not obvious, and how to obtain relatively accurate permeability is a key point and a difficulty in reservoir prediction.

[0003] The commonly used seismic permeability prediction methods mainly include the following: (1) Empirical formula: By extending the relationship between porosity and permeability in wells to seismic permeability prediction, it is difficult to accurately predict permeability when the relationship between porosity and permeability is complex. (2) Geostatistical simulation: Under the constraint of seismic attributes, the in-well permeability is interpolated and extrapolated to achieve permeability prediction. This method is difficult to apply under complex geological conditions and rapid lateral changes of reservoirs. Since there are certain limitations in using the above methods for permeability prediction, there is an urgent need for a permeability quantitative prediction method with strong adaptability. Summary of the Invention

[0004] Permeability is an important parameter for characterizing reservoir characteristics and fluid mobility, which can effectively guide oil and gas exploration and development. However, the permeability response characteristics in seismic data are not obvious, and conventional permeability prediction methods all have certain applicability and limitations. Based on this, the present invention comprehensively integrates multiple information such as seismic, horizon, and logging data, and proposes a method for quantitatively predicting permeability by integrating multiple means.

[0005] To achieve the above object, the present invention provides a method for quantitatively predicting permeability by integrating multiple means, including:

[0006] Obtain seismic data, horizon data, and logging data, and transform the seismic data, horizon data, and logging data into the same domain;

[0007] Based on the seismic data, horizon data, and logging data, respectively obtain the interpolated permeability perm_inter, the rock physics permeability perm_rp, and the machine learning permeability perm_ml;

[0008] Based on the obtained interpolation permeability perm_inter, petrophysical permeability perm_rp, and machine learning permeability perm_ml, calculate the final multi-method integrated permeability perm_final.

[0009] Further, the obtaining of the interpolation permeability perm_inter includes:

[0010] Extract various seismic attributes and screen out the seismic attributes sensitive to permeability;

[0011] Based on the comprehensive seismic attributes, horizons, and logging permeability results, obtain the interpolation permeability perm_inter by the method of inverse proportional weighting.

[0012] Further, the obtaining of the petrophysical permeability perm_rp includes:

[0013] Based on the petrophysical analysis results of logging data, select the elastic parameters or combinations of elastic parameters sensitive to permeability, and establish the conversion relationship between them and permeability;

[0014] Based on the seismic prestack inversion results, obtain the petrophysical permeability perm_rp.

[0015] Further, the obtaining of the machine learning permeability perm_ml includes:

[0016] Extract the seismic attributes of the well-side trace and the prestack inversion results, combine with the logging permeability to make a sample set, and obtain the machine learning permeability perm_ml based on the machine learning method.

[0017] Further, calculating the multi-method integrated permeability perm_final includes: obtaining the weights of the interpolation permeability perm_inter, petrophysical permeability perm_rp, and machine learning permeability perm_ml based on the ant colony algorithm, and then obtaining the final multi-method integrated permeability perm_final.

[0018] Further, extract various seismic attributes such as amplitude, frequency, and phase from seismic data, and quantitatively select the seismic attributes sensitive to permeability based on correlation analysis;

[0019] Use the evaluation function f to evaluate the non-linear relationship between seismic attributes and permeability, and its definition is as follows:

[0020] f = λ 1 *ρ 1 +λ 2 *ρ 2 (1)

[0021] Where: ρ 1is the Spearman rank correlation coefficient, ρ 2 is the Kendall rank correlation coefficient, λ 1 and λ 2 are weighting coefficients.

[0022] Furthermore, the multi - method integrated permeability perm_final is linearly combined by the interpolation permeability perm_inter, the petrophysical permeability perm_rp, and the machine - learning permeability perm_ml, that is:

[0023] perm_final = α1 * perm_inter+α2 * perm_rp+α3 * perm_ml (2)

[0024] where α1, α2, and α3 are weight coefficients, and the optimal weight coefficients are obtained based on the ant colony algorithm.

[0025] According to another aspect of the present invention, a device for quantitatively predicting the multi - method integrated permeability is provided, including:

[0026] A data acquisition module, which acquires seismic data, horizon data, and logging data, and transforms the seismic data, horizon data, and logging data into the same domain;

[0027] A permeability acquisition module, which respectively acquires the interpolation permeability perm_inter, the petrophysical permeability perm_rp, and the machine - learning permeability perm_ml based on the seismic data, horizon data, and logging data;

[0028] A permeability integration module, which calculates the final multi - method integrated permeability perm_final based on the acquired interpolation permeability perm_inter, petrophysical permeability perm_rp, and machine - learning permeability perm_ml.

[0029] According to another aspect of the present invention, an electronic device is provided, and the electronic device includes:

[0030] A memory, which stores executable instructions;

[0031] A processor, which runs the executable instructions in the memory to implement the multi - method integrated permeability quantitative prediction method.

[0032] According to another aspect of the present invention, a non - transitory computer - readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the multi - method integrated permeability quantitative prediction method is implemented.

[0033] The present invention has the following innovative points compared with the existing technologies:

[0034] (1) Based on the custom evaluation function f, the present invention can effectively screen out seismic attributes sensitive to permeability, providing a solid data basis for subsequent quantitative prediction of permeability.

[0035] (2) Utilizing the physical relationship between seismic data and permeability, the present invention obtains the mapping relationship between sensitive seismic attributes and the permeability curve based on the PINN network.

[0036] (3) Based on the ant colony algorithm, the present invention can obtain the optimal weight coefficients, and then linearly combine the interpolated permeability perm_inter, petrophysical permeability perm_rp, and machine learning permeability perm_ml to obtain the multi-method integrated permeability perm_final.

[0037] (4) The present invention establishes a process flow for the quantitative prediction of permeability by multi-method integration. In the process of quantitative prediction of permeability, the advantages of three permeability prediction methods are fully utilized, and the ant colony algorithm is introduced for organic integration to obtain accurate permeability prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] By describing the exemplary embodiments of the present invention in more detail in conjunction with the accompanying drawings, the above and other objects, features, and advantages of the present invention will become more apparent. Among them, in the exemplary embodiments of the present invention, the same reference numerals generally represent the same components.

[0039] Figure 1 It is a flow chart of the quantitative prediction method of permeability by multi-method integration according to the present invention.

[0040] Figure 2 It is a flow chart of the method implementation according to the embodiment of the present invention.

[0041] Figure 3 It is a calibration map of typical wells in the study area according to the embodiment of the present invention.

[0042] Figure 4 It is a predicted profile of interpolated permeability according to the embodiment of the present invention.

[0043] Figure 5 It is a predicted profile of multi-method integrated permeability according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] The preferred embodiments of the present invention will be described in more detail below. Although the preferred embodiments of the present invention are described below, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0045] The present invention belongs to the field of oil and gas geophysical exploration, and is a method for quantitatively predicting permeability by integrating multiple means. Permeability is an important parameter for characterizing reservoir characteristics and fluid mobility, which can effectively guide oil and gas exploration and development. However, the response characteristics of permeability in seismic data are not obvious, and conventional permeability prediction methods all have certain applicability and limitations.

[0046] In view of these problems, the present invention comprehensively utilizes multi-source information such as seismic, horizon, and logging data, and proposes a new method for quantitatively predicting permeability by integrating multiple means. First, based on the extraction of various seismic attributes, a seismic attribute sensitive to permeability is obtained through a custom evaluation function; then, three permeability prediction results are obtained based on well interpolation, rock physics, and machine learning methods respectively; finally, based on the ant colony algorithm, the optimal weights of the three permeability prediction results are obtained, and the final quantitative permeability prediction result is integrated. The application of examples shows that the permeability prediction result of this method is highly consistent with that of the well, and has great application and popularization value.

[0047] Embodiment 1

[0048] As Figure 1 shown, this embodiment provides a method for quantitatively predicting permeability by integrating multiple means, including:

[0049] Obtain seismic data, horizon data, and logging data, and transform the seismic data, horizon data, and logging data into the same domain;

[0050] Based on the seismic data, horizon data, and logging data, obtain the interpolated permeability perm_inter, the rock physics permeability perm_rp, and the machine learning permeability perm_ml respectively;

[0051] Based on the obtained interpolated permeability perm_inter, rock physics permeability perm_rp, and machine learning permeability perm_ml, calculate the final integrated permeability perm_final by multiple means.

[0052] Furthermore, the obtaining of the interpolated permeability perm_inter includes:

[0053] Extract various seismic attributes, and screen out the seismic attributes sensitive to permeability;

[0054] Based on the comprehensive seismic attributes, horizon, and logging permeability results, obtain the interpolated permeability perm_inter by the method of inverse proportional weighting.

[0055] Furthermore, the obtaining of the rock physics permeability perm_rp includes:

[0056] Based on the petrophysical analysis results of well logging data, select elastic parameters or combinations of elastic parameters that are sensitive to permeability, and establish the conversion relationship between them and permeability;

[0057] Based on the results of prestack seismic inversion, obtain the petrophysical permeability perm_rp.

[0058] Furthermore, the obtaining of the machine learning permeability perm_ml includes:

[0059] Extract seismic attributes of the well-side trace and prestack inversion results, combine with well logging permeability to make a sample set, and obtain the machine learning permeability perm_ml based on machine learning methods.

[0060] Furthermore, calculating the multi-method integrated permeability perm_final includes: obtaining the weights of the interpolation permeability perm_inter, petrophysical permeability perm_rp, and machine learning permeability perm_ml based on the ant colony algorithm, and then obtaining the final multi-method integrated permeability perm_final.

[0061] Furthermore, extract various seismic attributes such as amplitude, frequency, and phase from seismic data, and quantitatively select the seismic attributes sensitive to permeability based on correlation analysis;

[0062] Use the evaluation function f to evaluate the non-linear relationship between seismic attributes and permeability, and its definition is as follows:

[0063] f = λ 1 *ρ 1 +λ 2 *ρ 2 (1)

[0064] Where: ρ 1 is the Spearman rank correlation coefficient, ρ 2 is the Kendall rank correlation coefficient, λ 1 and λ 2 are weighting coefficients.

[0065] Furthermore, the multi-method integrated permeability perm_final is linearly combined by the interpolation permeability perm_inter, petrophysical permeability perm_rp, and machine learning permeability perm_ml, that is:

[0066] perm_final = α1 * perm_inter + α2 * perm_rp + α3 * perm_ml (2)

[0067] Where α1, α2, and α3 are weight coefficients. Based on the ant colony algorithm, the optimal weight coefficients can be obtained.

[0068] Example Two

[0069] As Figure 2 shown, in this example, by making full use of information such as seismic data, horizons, and well logs, a permeability prediction method integrating multiple means is established. The method flow is as follows:

[0070] Step 1: Data quality control and optimization to obtain optimized seismic, horizon, and well log data;

[0071] Step 2: Conduct fine well-seismic calibration to transform seismic, horizon, and well log information into the same domain;

[0072] Step 3: Seismic attribute extraction and optimization to extract various seismic attributes and screen out the seismic attributes sensitive to permeability;

[0073] Step 4: Obtain the interpolated permeability perm_inter. Based on the results of seismic attributes, horizons, and well log permeability, use the inverse proportional weighting method to obtain the interpolated permeability perm_inter;

[0074] Step 5: Based on the results of rock physics analysis, select the elastic parameters or combinations of elastic parameters sensitive to permeability, establish the conversion relationship between them and permeability, and then based on the results of prestack seismic inversion, obtain the rock physics permeability perm_rp;

[0075] Step 6: Extract the seismic attributes of the well-side trace and the results of prestack seismic inversion, combine with well log permeability to make a sample set, divide the training set and the validation set according to the ratio of 7:3, and obtain the machine learning permeability perm_ml based on the machine learning method;

[0076] Step 7: Use the interpolated permeability perm_inter, rock physics permeability perm_rp, and machine learning permeability perm_ml obtained above, based on the ant colony algorithm, obtain the optimal weights of the three permeability results, and then obtain the final multi-means integrated permeability perm_final.

[0077] Specifically, the multi-means integrated permeability quantitative prediction method of this example is carried out according to the following steps:

[0078] Step 1: Data quality control and optimization. It mainly includes removing outliers and standardizing seismic data, correcting well log curves and processing well-to-well consistency, etc., to obtain optimized seismic, horizon, and well log data.

[0079] Step 2: Fine well-seismic calibration. Use seismic, horizon, and well log data to conduct fine well-seismic calibration to transform seismic, horizon, and well log data into the same domain.

[0080] Step 3: Seismic attribute extraction and optimization. Extract various seismic attributes such as amplitude, frequency, and phase from seismic data, and quantitatively select the seismic attributes sensitive to permeability based on correlation analysis.

[0081] In this embodiment, an evaluation function f is defined to evaluate the non - linear relationship between seismic attributes and permeability, and its definition is as follows:

[0082] f = λ 1 *ρ 1 +λ 2 *ρ 2 (1)

[0083] Where: ρ 1 is the Spearman rank correlation coefficient, ρ 2 is the Kendall rank correlation coefficient, and λ 1 and λ 2 are weighting coefficients.

[0084] Step 4: Calculate the interpolated permeability perm_inter. Integrate seismic attributes, horizons, and well - log permeability results, and obtain the interpolated permeability perm_inter based on the inverse - proportion weighting method with the preferred sensitive attributes as constraints.

[0085] Step 5: Calculate the rock - physics permeability perm_rp. Combining the previous understanding, on the basis of obtaining various elastic parameters, calculate the fluid mobility, shear compliance factor, etc. sensitive to permeability, then through rock - physics analysis, screen the parameters sensitive to permeability, and then fit to obtain the conversion relationship between the sensitive parameters and permeability. Further, based on the prestack inversion results, obtain the rock - physics permeability perm_rp.

[0086] Step 6: Calculate the machine - learning permeability perm_ml. Extract the seismic attributes of the well - side trace, and combine with the predicted results of the well - log permeability to make a sample set. Use the physics - informed neural network (PINN) to obtain the mapping relationship between seismic attributes and reservoir parameters. PINN is a supervised neural network that adds physical - information constraints during training and can learn a model with good generalization ability through fewer samples.

[0087] Step 7: Calculate the multi - method integrated permeability perm_final. Use the ant - colony algorithm to obtain the optimal weight coefficients of the interpolated permeability perm_inter, the rock - physics permeability perm_rp, and the machine - learning permeability perm_ml, and perform a linear combination to obtain the multi - method integrated permeability perm_final.

[0088] The ant colony algorithm is an optimization algorithm that simulates the process of ants searching for the optimal path when looking for food. The walking paths of ants represent a solution to the problem to be optimized. The walking paths of all ants are the solution space of the problem. Ants with shorter paths will release more information. As time goes by, shorter paths will be chosen by more ants, and thus there will be more information. Eventually, all ants will gather on the optimal path, thereby obtaining the optimal solution.

[0089] It is assumed that the multi-method integrated permeability perm_final is linearly combined by the interpolated permeability perm_inter, the petrophysical permeability perm_rp, and the machine learning permeability perm_ml, that is:

[0090] perm_final = α1 * perm_inter + α2 * perm_rp + α3 * perm_ml (2)

[0091] Among them, α1, α2, and α3 are weight coefficients. Based on the ant colony algorithm, the optimal weight coefficients can be obtained, thereby obtaining the multi-method integrated permeability perm_final.

[0092] Example Three

[0093] Refer to Figures 3 - 5 As shown, an example is used to illustrate the implementation process and application effect of the present invention. Taking the data of a certain actual work area in China as an example, a quantitative prediction study of permeability is carried out.

[0094] Step 1: Data quality control and optimization. It mainly includes removing outliers and standardizing seismic data, correcting logging curves and processing well-to-well consistency, etc., to obtain optimized seismic, horizon, and logging data.

[0095] Step 2: Fine well-seismic calibration. Use seismic, horizon, and logging data to carry out fine well-seismic calibration, and convert seismic, horizon, and logging data into the same domain. Figure 3 This is the calibration result of a typical well in the study area.

[0096] Step 3: Seismic attribute extraction and optimization. Use seismic data to extract various seismic attributes such as amplitude, frequency, and phase, and quantitatively select the seismic attributes sensitive to permeability as the attenuation gradient, longitudinal wave impedance, and shear compliance factor based on a custom evaluation function.

[0097] Step 4: Calculate the interpolated permeability perm_inter. Based on the comprehensive seismic attributes, horizons, and logging permeability results, and taking the optimized sensitive attributes as constraints, obtain the interpolated permeability perm_inter based on the inverse proportional weighting method. Figure 4 This is the prediction result of the interpolated permeability perm_inter passing through Well A2.

[0098] Step 5: Calculate the petrophysical permeability perm_rp. Based on the previous understanding and on the basis of obtaining various elastic parameters, calculate the fluid mobility, shear compliance factor, etc. that are sensitive to permeability. Then, through petrophysical analysis, screen out the shear compliance factor that is more sensitive to permeability, and fit to obtain the conversion relationship between the shear compliance factor and permeability. Further, based on the pre-stack inversion results, obtain the petrophysical permeability perm_rp.

[0099] Step 6: Calculate the machine learning permeability perm_ml. Extract the attenuation gradient of the well-side trace, the longitudinal wave impedance, and the shear compliance factor, and combine with the predicted results of the in-well permeability to make a sample set. Use the physics-informed neural network (PINN) to obtain the mapping relationship between the sensitive attributes and the reservoir parameters, so as to obtain the machine learning permeability perm_ml.

[0100] Step 7: Calculate the multi-method integrated permeability perm_final. Use the ant colony algorithm to obtain the optimal weight coefficients of the interpolation permeability perm_inter, the petrophysical permeability perm_rp, and the machine learning permeability perm_ml, and perform a linear combination to obtain the multi-method integrated permeability perm_final, as Figure 5 shown. By comparing Figure 4 and Figure 5 it can be seen that the in-well coincidence degree of the predicted result of the multi-method integrated permeability is higher, which is better than the predicted result of the permeability obtained by interpolation alone, proving that this method has good application effects.

[0101] Example 4

[0102] This example provides a multi-method integrated permeability quantitative prediction device, including:

[0103] A data acquisition module that acquires seismic data, horizon data, and logging data, and converts the seismic data, horizon data, and logging data into the same domain;

[0104] A permeability acquisition module that respectively obtains the interpolation permeability perm_inter, the petrophysical permeability perm_rp, and the machine learning permeability perm_ml based on the seismic data, horizon data, and logging data;

[0105] A permeability integration module that calculates the final multi-method integrated permeability perm_final based on the obtained interpolation permeability perm_inter, petrophysical permeability perm_rp, and machine learning permeability perm_ml.

[0106] Specifically, the permeability acquisition module obtains the interpolation permeability perm_inter including:

[0107] Extract multiple seismic attributes and screen out the seismic attributes sensitive to permeability;

[0108] Based on the seismic attributes, horizons, and well log permeability results, obtain the interpolated permeability perm_inter using the inverse proportional weighting method.

[0109] The permeability acquisition module obtains the rock physics permeability perm_rp, including:

[0110] Based on the rock physics analysis results of well log data, select the elastic parameters or combinations of elastic parameters sensitive to permeability and establish the conversion relationship between them and permeability;

[0111] Based on the seismic pre-stack inversion results, obtain the rock physics permeability perm_rp.

[0112] The permeability acquisition module obtains the machine learning permeability perm_ml, including:

[0113] Extract the seismic attributes and pre-stack inversion results of the well-side traces, combine with the well log permeability to make a sample set, and obtain the machine learning permeability perm_ml based on the machine learning method.

[0114] The permeability integration module obtains the weights of the interpolated permeability perm_inter, rock physics permeability perm_rp, and machine learning permeability perm_ml based on the ant colony algorithm, and then obtains the final multi-method integrated permeability perm_final.

[0115] Embodiment 5

[0116] This embodiment provides an electronic device, which includes:

[0117] A memory storing executable instructions;

[0118] A processor that runs the executable instructions in the memory to implement the multi-method integrated permeability quantitative prediction method. The method includes:

[0119] Obtain seismic data, horizon data, and well log data, and transform the seismic data, horizon data, and well log data into the same domain;

[0120] Based on the seismic data, horizon data, and well log data, respectively obtain the interpolated permeability perm_inter, rock physics permeability perm_rp, and machine learning permeability perm_ml;

[0121] Based on the obtained interpolated permeability perm_inter, rock physics permeability perm_rp, and machine learning permeability perm_ml, calculate the final multi-method integrated permeability perm_final.

[0122] Example VI

[0123] This embodiment provides a non - transitory computer - readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the multi - means integrated permeability quantitative prediction method described above is implemented. The method includes:

[0124] Obtain seismic data, horizon data, and logging data, and transform the seismic data, horizon data, and logging data into the same domain;

[0125] Based on the seismic data, horizon data, and logging data, respectively obtain the interpolated permeability perm_inter, petrophysical permeability perm_rp, and machine - learning permeability perm_ml;

[0126] Based on the obtained interpolated permeability perm_inter, petrophysical permeability perm_rp, and machine - learning permeability perm_ml, calculate the final multi - means integrated permeability perm_final.

[0127] The above - mentioned computer - readable storage medium includes but is not limited to: optical storage media (such as CD - ROM and DVD), magneto - optical storage media (such as MO), magnetic storage media (such as magnetic tapes or external hard drives), media with built - in rewritable non - volatile memory (such as memory cards), and media with built - in ROM (such as ROM cartridges).

[0128] In summary, the present invention comprehensively utilizes multi - source information of seismic, horizon, and logging, and proposes a new multi - means integrated permeability quantitative prediction method. First, based on the extraction of various seismic attributes, seismic attributes sensitive to permeability are obtained through a custom - defined evaluation function; then, three permeability prediction results are obtained respectively based on well interpolation, petrophysics, and machine - learning methods; finally, based on the ant colony algorithm, the optimal weights of the three permeability prediction results are obtained, and the final permeability quantitative prediction result is integrated. The application of the example shows that the permeability prediction result of this method has a high degree of coincidence with the well data, and has great application and promotion value.

[0129] The embodiments of the present invention have been described above. The above description is exemplary and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art in the technical field without departing from the scope and spirit of the described embodiments.

Claims

1. A quantitative prediction method for permeability integrating multiple means, characterized in that, it includes: Obtain seismic data, horizon data and logging data, and transform the seismic data, horizon data and logging data into the same domain; Based on the seismic data, horizon data and logging data, respectively obtain the interpolated permeability perm_inter, petrophysical permeability perm_rp and machine learning permeability perm_ml; Based on the obtained interpolated permeability perm_inter, petrophysical permeability perm_rp and machine learning permeability perm_ml, calculate the final multi - means integrated permeability perm_final.

2. The quantitative prediction method for permeability integrating multiple means according to claim 1, characterized in that, The obtaining of the interpolated permeability perm_inter includes: Extract multiple seismic attributes and screen out the seismic attributes sensitive to permeability; Based on the comprehensive seismic attributes, horizons and logging permeability results, obtain the interpolated permeability perm_inter based on the inverse proportional weighting method.

3. The quantitative prediction method for permeability integrating multiple means according to claim 1, characterized in that, The obtaining of the petrophysical permeability perm_rp includes: Based on the petrophysical analysis results of logging data, select the elastic parameters or combinations of elastic parameters sensitive to permeability, and establish the conversion relationship between them and permeability; Based on the prestack seismic inversion results, obtain the petrophysical permeability perm_rp.

4. The quantitative prediction method for permeability integrating multiple means according to claim 1, characterized in that, The obtaining of the machine learning permeability perm_ml includes: Extract the seismic attributes of the well - side trace and the prestack seismic inversion results, combine with the logging permeability to make a sample set, and obtain the machine learning permeability perm_ml based on the machine learning method.

5. The quantitative prediction method for permeability integrating multiple means according to claim 1, characterized in that, Calculating the multi - means integrated permeability perm_final includes: obtaining the weights of the interpolated permeability perm_inter, petrophysical permeability perm_rp and machine learning permeability perm_ml based on the ant colony algorithm, and then obtaining the final multi - means integrated permeability perm_final.

6. The quantitative prediction method for permeability integrating multiple means according to claim 4, characterized in that, Extract various seismic attributes such as amplitude, frequency and phase from seismic data, and quantitatively select the seismic attributes sensitive to permeability based on correlation analysis; Use the evaluation function f to evaluate the non - linear relationship between seismic attributes and permeability, and its definition is as follows: f = λ 1 *ρ 1 + λ 2 *ρ 2 (1) where: ρ 1 is the Spearman rank correlation coefficient, ρ 2 is the Kendall rank correlation coefficient, λ 1 and λ 2 are the weighting coefficients.

7. The quantitative prediction method for permeability integrating multiple means according to claim 5, characterized in that, The multi - means integrated permeability perm_final is linearly combined by the interpolated permeability perm_inter, petrophysical permeability perm_rp and machine learning permeability perm_ml, that is: perm_final = α1 * perm_inter + α2 * perm_rp + α3 * perm_ml (2) Wherein, α1, α2, and α3 are weight coefficients, and the optimal weight coefficients are obtained based on the ant colony algorithm.

8. A quantitative permeability prediction device integrating multiple means Characterized in that It includes A data acquisition module, which acquires seismic data, horizon data, and logging data, and converts the seismic data, horizon data, and logging data into the same domain A permeability acquisition module, which respectively acquires the interpolated permeability perm_inter, the rock physics permeability perm_rp, and the machine learning permeability perm_ml based on the seismic data, horizon data, and logging data A permeability integration module, which calculates the final multi - means integrated permeability perm_final based on the acquired interpolated permeability perm_inter, rock physics permeability perm_rp, and machine learning permeability perm_ml 9. An electronic device Characterized in that The electronic device includes A memory, which stores executable instructions A processor, which runs the executable instructions in the memory to implement the multi - means integrated quantitative permeability prediction method according to any one of claims 1 - 7 10. A non - transitory computer - readable storage medium, on which a computer program is stored Characterized in that The computer program, when executed by a processor, implements the multi - means integrated quantitative permeability prediction method according to any one of claims 1 - 7

Images