System and method for quantitative prediction of lamellar fracture development distribution

Abstract

The invention relates to the technical field of oil and gas field exploration, in particular to a lamellar fracture development distribution quantitative prediction system and method. The system comprises a data acquisition module, a development intensity evaluation module, a feature extraction module, a processor, a data processing module and a page arrangement crack prediction module, the data acquisition module is connected with the processor through a signal, and the data processing module is connected with the processor through a signal. And the processor is connected with the development intensity evaluation module, the feature extraction module and the lapage seam prediction module through signals. According to the system and the method provided by the invention, the main control factor-distribution trend-development intensity relationship is constructed, and the lamellar fracture prediction model is constructed based on the main control factor-distribution trend-development intensity relationship, so that visual prediction of the lamellar fracture development intensity and distribution trend is realized, the lamellar fracture development distribution accuracy is improved, and the lamellar fracture prediction efficiency is improved. And compared with the existing prediction means, the method is more visual.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas field exploration, and particularly to a system and method for quantitatively predicting the development and distribution of bedding fissures. Background Art

[0002] Bedding fissures mainly refer to the fissures formed during sedimentation and diagenesis, along or parallel to the bedding plane. In practical applications, shale will discharge a large amount of original formation water through diagenetic compaction. The pore radius of shale formed by fine particles is extremely small, and groundwater is difficult to pass freely through it. Therefore, it is difficult to obtain the cementing substances brought by the dissolution of formation water between bedding planes. The uncemented bedding laminae become fragile surfaces. Once there is a concentration of tectonic stress or the volume expansion of a large amount of hydrocarbon substances generated in the shale, it is extremely easy to induce the formation of fissures between the bedding planes, that is, bedding fissures.

[0003] As an effective channel for shale oil and gas seepage, bedding fissures enable dense shale reservoirs to also become effective producing layers, and relatively high oil and gas production capacity can occur. Therefore, the distribution law of bedding fissures is an important geological basis for shale oil and gas exploration and development, and is of great significance for shale oil and gas exploration and development.

[0004] For example, in a Chinese patent with the publication number CN110850502B, a method, device and system for predicting the distribution law of bedding fissures in shale oil and gas reservoirs are disclosed. The method includes identifying the developed sections of bedding fissures in a single-well shale oil and gas reservoir; evaluating the development intensity of the identified bedding fissures in the single-well shale oil and gas reservoir to obtain the development intensity of the bedding fissures in the single well; clarifying the main controlling factors of the bedding fissures and establishing the relationship information between them and the development intensity of the bedding fissures in the single well; and predicting the distribution law of bedding fissures in the shale oil and gas reservoir according to the development intensity of the bedding fissures in the single well and the relationship information.

[0005] This method identifies the developed sections of bedding fissures in a single-well shale oil and gas reservoir and evaluates the development intensity, clarifies the main controlling factors for the development of bedding fissures in the shale oil and gas reservoir, and establishes the relationship between each main controlling factor of the bedding fissures and the degree of development of the bedding fissures to predict the development of the bedding fissures. However, this method can only obtain the intensity of the bedding fissures and cannot predict the bedding fissure trend, resulting in the problem that the prediction of bedding fissures is not comprehensive enough. Therefore, in view of the above deficiencies, a system and method for quantitatively predicting the development and distribution of bedding fissures are proposed. Summary of the Invention

[0006] (I) Technical Problems to be Solved

[0007] The present invention provides a system and method for quantitatively predicting the development and distribution of bedding fissures to overcome the problem that the existing technology can only obtain the intensity of the bedding fissures and cannot predict the bedding fissure trend, resulting in the problem that the prediction of bedding fissures is not comprehensive enough.

[0008] (II) Technical Solutions

[0009] To solve the above problems, the present invention provides a quantitative prediction system for the development and distribution of bedding plane joints, including:

[0010] A data acquisition module, a development intensity evaluation module, a feature extraction module, a processor, a data processing module, and a bedding plane joint prediction module. The data acquisition module is connected to the processor through a signal, the data processing module is connected to the processor through a signal, and the processor is respectively connected to the development intensity evaluation module, the feature extraction module, and the bedding plane joint prediction module through signals.

[0011] Preferably, the data acquisition module is used to obtain the development information data of the drilling cores and thin sections in the area to be measured; the development intensity evaluation module is used to determine the development intensity of the drilling cores and thin sections in the area to be measured; the feature extraction module is used to extract the main control factors among the influencing factors of the bedding plane joints in the area to be measured; the processor is used for the construction and optimization of the prediction model; the data processing module is used to establish the relationship between the main control factors, the distribution trend, and the development intensity; and the bedding plane joint prediction module is used to quantitatively and directionally predict the distribution and development intensity of the bedding plane joints based on the bedding plane joint prediction model.

[0012] The present invention also provides a method based on the quantitative prediction system for the development and distribution of bedding plane joints, including:

[0013] Step S1: Determine the area to be measured, obtain well logging data and drilling cores through the data acquisition module, make the drilling cores into thin sections, and determine the development intensity of the bedding plane joints through the development intensity evaluation module according to the development data of the drilling cores and thin sections.

[0014] Step S2: Determine the main control factors in the area to be measured by using the feature extraction module according to the well logging data and development data obtained in Step S1.

[0015] Step S3: Construct a main control factor - distribution trend - development intensity relationship function by using the data processing module according to the main control factors determined in Step S2 and the development intensity data determined in Step S1.

[0016] Step S4: According to the relationship function obtained in Step S3, adopt the Monte Carlo method to construct a pre-training model through the processor.

[0017] Step S5: Optimize the pre-training model established in Step S4 through the processor to obtain a bedding plane joint prediction model.

[0018] Step S6: Carry out quantitative prediction of the development and distribution of bedding plane joints in the area to be measured according to the bedding plane joint prediction model obtained in Step S5.

[0019] Preferably, in the step S1, the logging data includes acoustic travel time, natural gamma, spontaneous potential, and resistivity.

[0020] Preferably, in the step S2, the main controlling factors include mineral composition, TOC content, shale thickness, rock mechanical properties, and bedding distribution characteristics.

[0021] Preferably, in the step S3, the calculation formula for the main controlling factor - distribution trend - development intensity relationship function constructed by the data processing module is:

[0022]

[0023] where SFI is the development intensity of the bedding fissures, F(x,y) is the distribution trend of the center points of the bedding fissures, w i is the weight of the i-th main controlling factor, f i (FI) is the relationship between the i-th main controlling factor and the development intensity of single-section bedding fissures, f i (x i ,y j ) is the relationship between the i-th main controlling factor and the distribution trend of single-section bedding fissures.

[0024] Preferably, in the step S4, the processor constructs the pre-training model as follows:

[0025] S41. Based on the development data of drilling cores and thin sections, construct a Monte Carlo prediction model;

[0026] S42. Based on the development data of drilling cores and thin sections, use the Monte Carlo method to obtain simulation data;

[0027] S43. Substitute the simulation data and the main controlling factor - distribution trend - development intensity relationship function into the Monte Carlo prediction model for training to obtain the pre-training model.

[0028] Preferably, in the step S5, the method for the processor to optimize the pre-training model is the maximum mean discrepancy method.

[0029] (III) Beneficial effects

[0030] The quantitative prediction system and method for the development and distribution of bedding fissures provided by the present invention realizes the visual prediction of the development intensity and distribution trend of bedding fissures by constructing the main controlling factor - distribution trend - development intensity relationship, constructs a bedding fissure prediction model based on the main controlling factor - distribution trend - development intensity relationship, improves the accuracy of the development and distribution of bedding fissures, and is more intuitive compared with the existing prediction means. Description of the drawings

[0031] Figure 1Schematic diagram of modules of the quantitative prediction system for bedding joint development and distribution according to an embodiment of the present invention;

[0032] Figure 2 Schematic flow chart of the method of the quantitative prediction system for bedding joint development and distribution according to an embodiment of the present invention. Detailed implementation manners

[0033] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0034] Figure 1 Schematic diagram of modules of the quantitative prediction system for bedding joint development and distribution according to an embodiment of the present invention, as Figure 1 shown, the present invention proposes a quantitative prediction system for bedding joint development and distribution, specifically including:

[0035] A data acquisition module, a development intensity evaluation module, a feature extraction module, a processor, a data processing module, and a bedding joint prediction module. The data acquisition module is signal-connected to the processor, the data processing module is signal-connected to the processor, and the processor is signal-connected to the development intensity evaluation module, the feature extraction module, and the bedding joint prediction module respectively.

[0036] In practical applications, the data acquisition module is used to obtain the development information data of the drilling cores and thin sections in the area to be measured; the development intensity evaluation module is used to determine the development intensity of the drilling cores and thin sections in the area to be measured; the feature extraction module is used to extract the main control factors among the factors affecting the bedding joints in the area to be measured; the processor is used to construct and optimize the prediction model; the data processing module is used to establish the relationship between the main control factors and the distribution trend and development intensity; the bedding joint prediction module is used to quantitatively and directionally predict the distribution and development intensity of the bedding joints based on the bedding joint prediction model.

[0037] Figure 2 Schematic flow chart of the method of the quantitative prediction system for bedding joint development and distribution according to an embodiment of the present invention, as Figure 2 shown, the present invention also provides a method based on the quantitative prediction system for bedding joint development and distribution, including:

[0038] Step S1: Determine the area to be measured, obtain well logging data and drilling cores through the data acquisition module, make the drilling cores into thin sections, and determine the development intensity of the bedding joints through the development intensity evaluation module according to the development data of the drilling cores and thin sections;

[0039] Step S2: Based on the logging data and development data obtained in Step S1, use the feature extraction module to determine the main controlling factors of the area to be measured;

[0040] Step S3: Based on the main controlling factors determined in Step S2 and the development intensity data determined in Step S1, use the data processing module to construct a main controlling factor - distribution trend - development intensity relationship function;

[0041] Step S4: Based on the degree relationship function obtained in Step S3, use the Monte Carlo method to construct a pre-training model through the processor;

[0042] Step S5: Optimize the pre-training model established in Step S4 through the processor to obtain a bedding fracture prediction model;

[0043] Step S6: Based on the bedding fracture prediction model obtained in Step S5, conduct a quantitative prediction of the development distribution of bedding fractures in the area to be measured.

[0044] In this prediction method, in Step S1, the logging data includes acoustic travel time, natural gamma, spontaneous potential, and resistivity; in Step S2, the main controlling factors include mineral composition, TOC content, shale thickness, rock mechanical properties, and lamination distribution characteristics.

[0045] In practical applications, in Step S3, the calculation formula for the data processing module to construct the main controlling factor - distribution trend - development intensity relationship function is:

[0046]

[0047] Among them, SFI is the development intensity of bedding fractures, F(x, y) is the distribution trend of the center points of bedding fractures, w i is the weight of the i-th main controlling factor, f i (FI) is the relationship between the i-th main controlling factor and the development intensity of a single segment of bedding fractures, f i (x i , y j ) is the relationship between the i-th main controlling factor and the distribution trend of a single segment of bedding fractures.

[0048] In this prediction method, in Step S4, the specific steps for the processor to construct the pre-training model are as follows:

[0049] S41. Based on the development data of drilling cores and thin sections, construct a Monte Carlo prediction model;

[0050] S42. Based on the development data of drilling cores and thin sections, use the Monte Carlo method to obtain simulation data;

[0051] S43. Substitute the simulation data and the main controlling factor - distribution trend - development intensity relationship function into the Monte Carlo prediction model for training to obtain the pre-training model.

[0052] In practical applications, the Monte Carlo method is used to construct a prediction model, which requires fewer sample numbers, reducing the cost loss caused by obtaining the measured development data of drilling cores and thin sections. At the same time, the prediction model is optimized based on the maximum mean discrepancy method, improving the generalization ability and robustness of the prediction model, and further enhancing the accuracy of the bedding joint prediction model in predicting the development distribution of bedding joints.

[0053] It should be noted that in step S5, the method for the processor to optimize the pre-trained model is the maximum mean discrepancy method.

[0054] The following specifically describes the specific working principle of this quantitative prediction system and method based on the development distribution of bedding joints:

[0055] In this embodiment, a prediction system is constructed by signal-connecting a data acquisition module, a development intensity evaluation module, a feature extraction module, a processor, a data processing module, and a bedding joint prediction module;

[0056] Among them, the data acquisition module is used to obtain the development information data of the drilling cores and thin sections in the area to be measured;

[0057] The development intensity evaluation module is used to determine the development intensity of the drilling cores and thin sections in the area to be measured;

[0058] The feature extraction module is used to extract the main control factors among the influencing factors of bedding joints in the area to be measured;

[0059] The processor is used for the construction and optimization of the prediction model;

[0060] The data processing module is used to establish the relationship between the main control factors and the distribution trend and development intensity;

[0061] The bedding joint prediction module is used to quantitatively and directionally predict the distribution and development intensity of bedding joints based on the bedding joint prediction model.

[0062] In practical applications, the specific implementation process of the method based on the quantitative prediction system of the development distribution of bedding joints is as follows:

[0063] Step 1: Determine the area to be measured, obtain well logging data and drilling cores, make the drilling cores into thin sections, and based on the development data of the drilling cores and thin sections, the development intensity evaluation module determines the development intensity of the bedding joints.

[0064] In this embodiment, the well logging data includes acoustic time difference, natural gamma, spontaneous potential, and resistivity.

[0065] Step 2: According to the well logging data and development data in the data acquisition module, the feature extraction module determines the main control factors in the area to be measured.

[0066] In this embodiment, the main control factors include mineral composition, TOC content, shale thickness, rock mechanical properties, and bedding distribution characteristics.

[0067] Step 3: According to the main control factors determined by the feature extraction module and the development intensity data determined by the development intensity evaluation module, use the data processing module to construct a main control factor - distribution trend - development intensity relationship function.

[0068] In this embodiment, the steps for the data processing module to construct the main control factor - distribution trend - development intensity relationship function are as follows:

[0069] S31: Evaluate the development intensity FI of single - segment bedding fractures:

[0070]

[0071] S32: Evaluate the distribution trend of single - segment bedding fractures:

[0072] f(x i ,y j )=f(x i-1 +a,y j-1 +b),

[0073]

[0074] where x i and y j are respectively the horizontal coordinate and the vertical coordinate of the center point of the single - segment bedding fracture, x i-1 and y j-1 are respectively the horizontal coordinate and the vertical coordinate of the center point of the previous - segment bedding fracture, a is the horizontal change distance between the center points of adjacent bedding fractures, b is the vertical change distance between the center points of adjacent bedding fractures, α is the horizontal change inclination angle of the center points of adjacent bedding fractures, and θ is the vertical change inclination angle of the center points of adjacent bedding fractures;

[0075] S33: Based on the main control factors, construct the main control factor - distribution trend - development intensity relationship function:

[0076]

[0077] where SFI is the development intensity of the bedding fracture, F(x, y) is the distribution trend of the center point of the bedding fracture, w i is the weight of the i - th main control factor, f i (FI) is the relationship between the i - th main control factor and the development intensity of the single - segment bedding fracture, and f i (x i ,y j ) is the relationship between the i - th main control factor and the distribution trend of the single - segment bedding fracture.

[0078] Step 4: According to the master factor - distribution trend - development intensity relationship function, use the Monte Carlo method to construct a pre-training model through a processor.

[0079] In this embodiment, the specific steps for the processor to construct a pre-training model through the Monte Carlo method are as follows:

[0080] S41: Based on the development data of drilling cores and thin sections, construct a Monte Carlo prediction model;

[0081] S42: Based on the development data of drilling cores and thin sections, use the Monte Carlo method to obtain simulation data;

[0082] S43: Substitute the simulation data and the master factor - distribution trend - development intensity relationship function into the Monte Carlo prediction model for training to obtain a pre-training model.

[0083] Step 5: Optimize the pre-training model through a processor to obtain a bedding joint prediction model.

[0084] In this embodiment, the processor optimizes the pre-training model according to the maximum mean discrepancy method to obtain a bedding joint prediction model. The specific process is as follows:

[0085] S51: Let L mmd be the square distance of the maximum mean discrepancy, F s , F t be the features extracted from the measured data and simulation data of the development data, M and N be the sample numbers of the measured data and simulation data, Φ be the kernel function for mapping the original data to the reproducing Hilbert space, be the samples of the measured data and simulation data, and calculate the square distance of the maximum mean discrepancy L mmd between the simulation data and the measured data. The formula is as follows:

[0086]

[0087] S52: Add the square distance of the maximum mean discrepancy to the cross-entropy loss function of the pre-training model to obtain the total loss function L total , as follows:

[0088]

[0089] where L total is the total loss function, X t , Y t are the features of the measured samples;

[0090] S53: Perform transfer learning based on the total loss function L total to obtain a bedding joint prediction model.

[0091] Step 6: According to the obtained bedding joint prediction model, quantitatively predict the development and distribution of bedding joints in the area to be measured.

[0092] In this embodiment, according to the obtained bedding joint prediction model, not only the development intensity of bedding joints can be predicted, but also the distribution trend of bedding joints can be visually predicted.

[0093] The quantitative prediction system and method for the development and distribution of bedding joints provided by the present invention construct the relationship of main control factors - distribution trend - development intensity, and based on this relationship, a bedding joint prediction model is constructed, realizing the visual prediction of the development intensity and distribution trend of bedding joints, improving the accuracy of the development and distribution of bedding joints, and being more intuitive compared with the existing prediction means.

[0094] The above embodiments are only used to illustrate the present invention, rather than limiting the present invention. Those of ordinary skill in the relevant technical fields can also make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, all equivalent technical solutions also belong to the scope of the present invention, and the patent protection scope of the present invention shall be defined by the claims.

Claims

1. A quantitative prediction system for the development and distribution of bedding plane joints, characterized in that, Including: A data acquisition module, a development intensity evaluation module, a feature extraction module, a processor, a data processing module, and a bedding joint prediction module. The data acquisition module is signal-connected to the processor, the data processing module is signal-connected to the processor, and the processor is respectively signal-connected to the development intensity evaluation module, the feature extraction module, and the bedding joint prediction module.

2. The quantitative prediction system for bedding joint development and distribution according to claim 1, wherein The data acquisition module is used to obtain the development information data of the drilling cores and thin sections in the area to be measured; the development intensity evaluation module is used to determine the development intensity of the drilling cores and thin sections in the area to be measured; the feature extraction module is used to extract the main control factors among the influencing factors of the bedding joints in the area to be measured; the processor is used for the construction and optimization of the prediction model; the data processing module is used to establish the relationship between the main control factors, the distribution trend, and the development intensity; the bedding joint prediction module is used to quantitatively and directionally predict the distribution and development intensity of the bedding joints based on the bedding joint prediction model.

3. A method for a quantitative prediction system of bedding joint development distribution proposed based on any one of claims 1 to 2, characterized in that, Including: Step S1: Determine the area to be measured, obtain well logging data and drilling cores through the data acquisition module, make the drilling cores into thin sections, and determine the development intensity of the bedding joints through the development intensity evaluation module according to the development data of the drilling cores and thin sections. Step S2: Determine the main control factors in the area to be measured by using the feature extraction module according to the well logging data and development data obtained in Step S1. Step S3: Construct a main control factor - distribution trend - development intensity relationship function by using the data processing module according to the main control factors determined in Step S2 and the development intensity data determined in Step S1. Step S4: According to the relationship function of degree obtained in Step S3, adopt the Monte Carlo method to construct a pre-training model through the processor. Step S5: Optimize the pre-training model established in Step S4 through the processor to obtain a bedding joint prediction model. Step S6: Carry out quantitative prediction of the development distribution of bedding joints in the area to be measured according to the bedding joint prediction model obtained in Step S5.

4. The method for the quantitative prediction system of bedding plane joint development and distribution according to claim 3, wherein In Step S1, the well logging data includes acoustic time difference, natural gamma, spontaneous potential, and resistivity.

5. The method of the quantitative prediction system for the development and distribution of bedding joints according to claim 3, characterized in that, In Step S2, the main control factors include mineral composition, TOC content, shale thickness, rock mechanical properties, and lamination distribution characteristics.

6. The method of the quantitative prediction system for bedding joint development and distribution according to claim 3, wherein In Step S3, the calculation formula for the data processing module to construct the main control factor - distribution trend - development intensity relationship function is: Among them, SFI is the development intensity of bedding joints, F(x, y) is the distribution trend of the center points of bedding joints, w i is the weight of the i-th main control factor, f i (FI) is the relationship between the i-th main control factor and the development intensity of single-section bedding joints, f i (x i , y j ) is the relationship between the i-th main control factor and the distribution trend of single-section bedding joints.

7. The method of the quantitative prediction system for bedding joint development and distribution according to claim 3, wherein In Step S4, the specific process for the processor to construct the pre-training model is as follows: S41. Based on the development data of the drilling cores and thin sections, construct a Monte Carlo prediction model. S42. Based on the development data of the drilling cores and thin sections, obtain simulation data by using the Monte Carlo method. S43. Substitute the simulation data and the main control factor - distribution trend - development intensity relationship function into the Monte Carlo prediction model for training to obtain a pre-training model.

8. The method of the quantitative prediction system for the development and distribution of bedding joints according to claim 3, wherein, In Step S5, the method for the processor to optimize the pre-training model is the maximum mean discrepancy method.

Images