A shale gas reservoir drilling lithofacies identification method, system and device

Abstract

The application discloses a shale gas reservoir rock facies identification method, system and equipment, and belongs to the technical field of shale gas exploration engineering, and comprises the following steps: obtaining shale gas reservoir core sample data and element logging data, and performing dimensionless processing on the element logging data; based on the core sample data, establishing a shale gas reservoir rock facies classification standard, and classifying shale gas reservoir rock facies types; based on the element logging data, screening target data through grey correlation analysis; and based on the shale gas reservoir rock facies types and the target data, obtaining target shale gas reservoir rock facies identification information. According to the application, the shale gas reservoir core sample data and the element logging data are used to classify rock facies types, and target data associated with rock facies in the element logging data is screened, and then the rock facies identification module is used to realize target shale gas reservoir rock facies identification, so that the technical problem of low timeliness of rock facies identification in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of shale gas exploration engineering technology, and in particular to a method, system and equipment for identifying shale gas reservoir facies while drilling. Background Technology

[0002] Fine-grained lithofacies classification of shale gas reservoirs is of great significance for deepening our understanding of sedimentary environments, clarifying enrichment patterns, and assessing development potential. It serves as a key link between geological theory and engineering practice. Currently, the most commonly used method is lithofacies classification based on core analysis. However, this method is costly, highly dependent on personal experience, and struggles to efficiently handle large volumes of data and complex, ever-changing geological conditions. It also lacks automation and objectivity. As the pace of shale gas exploration and development accelerates, improving timeliness while ensuring identification accuracy has become a critical challenge. Existing technologies cannot acquire data during drilling, resulting in poor timeliness. Furthermore, conventional logging data has limited capacity to characterize lithofacies information, leading to low accuracy. Summary of the Invention

[0003] The purpose of this invention is to address the shortcomings of existing technologies, such as the difficulty in acquiring lithofacies data in real time during drilling and the low accuracy of the acquired lithofacies data, and to propose a method, system, and equipment for lithofacies identification in shale gas reservoirs during drilling.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: S1. Obtain core sample data and elemental logging data of shale gas reservoirs, and perform dimensionless processing on the elemental logging data; The core sample data is based on the database of the target shale gas reservoir, historical data or experimental analysis, and the core sample data includes core sample X-ray diffraction data and core sample organic carbon content data. S2. Based on the core sample data, establish a shale gas reservoir lithofacies classification standard to classify shale gas reservoir lithofacies types; S3. Based on the dimensionless element logging data, target data is filtered through grey relational analysis. S4. Based on the core facies type and target data, obtain drilling facies identification information for the target shale gas reservoir.

[0005] In some feasible embodiments, the specific method for performing dimensionless processing on the elemental logging data is as follows: ; in, The data consists of elemental logging data after dimensionless processing. For elemental logging data, This represents the maximum content of the same element across all samples. This represents the minimum content of the same element in all samples.

[0006] In some feasible embodiments, S2 specifically includes the following steps: S21. Based on the organic carbon content data of core samples, set an organic carbon content threshold, and classify the core samples into three organic matter lithofacies according to the organic matter abundance based on the organic carbon content threshold. Based on the organic carbon content data of the core samples, 2% and 3% were used as the boundary for organic matter abundance classification, and an organic carbon content threshold was set. When the organic carbon content in the core sample is greater than 3%, it is classified as an organic-rich rock facies; when the organic carbon content in the core sample is in the range of 2% to 3%, it is classified as a medium-organic rock facies; and when the organic carbon content in the core sample is less than 2%, it is classified as an organic-poor rock facies. S22. Based on X-ray diffraction data of core samples, shale gas reservoirs are classified into 10 different types of shale gas reservoir lithology. Based on the different contents of felsic minerals, carbonate minerals, and clay minerals in the core samples, shale gas reservoirs are classified into 10 different shale gas reservoir lithologies. These 10 different shale gas reservoir lithologies include high-siliceous shale, high-calcareous shale, high-muddy shale, medium-siliceous shale, medium-calcareous shale, medium-muddy shale, siliceous and calcareous mixed shale, calcareous and muddy mixed shale, siliceous and muddy mixed shale, and impurity mixed shale. S23. Draw a three-terminal element diagram based on 10 different types of shale gas reservoir lithology; A three-endpoint diagram was drawn using the contents of felsic minerals, carbonate minerals, and clay minerals as the three endpoints. S24. Couple the organic matter abundance classification results with lithological types to establish a shale gas reservoir lithological classification standard and classify shale gas reservoir lithological types. By combining 10 different shale gas reservoir lithologies with 3 organic matter facies, 30 comprehensive facies categories of shale gas reservoirs were obtained.

[0007] In some feasible implementation methods, the screening of target data through grey relational analysis specifically includes the following steps: S31. Based on the dimensionless elemental logging data, obtain the grey relational coefficient between the core lithofacies and the elemental logging data. The specific method is as follows: The core facies sorted by depth after classification are used as a reference sequence: ; in, This is a reference sequence of core facies. For the first The core lithofacies values ​​corresponding to each sample point The index of the data in the sequence. The sequence length is the total number of data points.

[0008] Elemental logging data of the same depth, after being processed without dimension, are sorted by depth as a comparison sequence: ; in, For the first The dimensionless element logging comparison sequence corresponding to each element. For the first The element in the first Dimensionless element logging values ​​at each sample point For element type serial number, The number of element types; Calculate the grey relational coefficients between the reference sequence and the comparison sequence: ; in, For the first The element in the first Grey relational coefficient at each sample point This is the two-level minimum absolute difference between the reference sequence corresponding to all elements and all sample points and the sequence of the object to be evaluated. This is the two-level maximum absolute difference between the reference sequence corresponding to all elements and all sample points and the sequence of the object to be evaluated. The resolution coefficient is set to 0.5 based on database, historical data, or experimental analysis. S32. Based on the grey relational coefficient, calculate the grey relational degree between element logging data and the overall grey relational degree between core lithofacies; The specific method is as follows: ; in, To compare the grey relational degree between the sequence and the reference sequence, if the calculated The higher the value, the higher the correlation between the element logging data and the core lithofacies. S33. Set a gray relational degree selection threshold, and select target data from several dimensionless element logging data based on the selection threshold; The selection threshold is set according to the geological characteristics and data quality of different well areas, and based on the set selection threshold, several types of logging data with a value greater than the selection threshold are selected as target data from the overall gray relational degree.

[0009] In some feasible embodiments, S4 includes the following steps: S41. Construct a core sample training dataset based on core lithofacies type and target data, specifically as follows: ; in, For core sample training dataset, This refers to the number of core samples. For the first The target data feature vector corresponding to each core sample For the first The lithological type label of the shale gas reservoir corresponding to each core sample. This indicates that the core samples are arranged by serial number. From 1 to A set of samples arranged sequentially; S42. Based on the core sample training dataset, construct a classification model to classify the lithofacies of the core samples; Based on the training data of each core sample in the core sample training dataset, a support vector machine (SVM) classification model is constructed. The specific method is as follows: ; in, For the weight vector, This is the offset. The set of relaxation variables corresponding to all core samples. This represents finding the minimum value over the weight vector, the offset, and the set of relaxation variables corresponding to all core samples. This is the penalty parameter, used to balance maximizing the classification margin and minimizing the training error. For the first Relaxation variables corresponding to each core sample Indicates all The relaxation variables of each core sample are summed. To mark the constraints, For weight transpose, It is a high-dimensional mapping function; where slack variables and constraints are set through databases, historical data, or experimental analysis; S43. Based on the classification model, obtain drilling facies identification information of the target shale gas reservoir; By constructing a decision function, the dimensionless elemental logging data of the target shale gas reservoir is input into the classification model, and the lithofacies identification information of the target shale gas reservoir is output through the decision function. The decision function is: ; in, This is a feature vector for identifying the lithofacies information of the target shale gas reservoir. This is a function for determining the sign of a sign. To support the set of feature vectors for the target data, This indicates that the summation is performed only on samples within the support vector set. For optimal Lagrange multipliers, This is the optimal offset; if If the value is greater than 0, then output the identification information of a certain type of lithofacies; if... If the value is less than 0, other types of rock facies identification information will be output.

[0010] In some feasible implementation methods, the construction of the classification model based on the core sample training dataset further includes: S421. Obtain the nonlinear distribution of the target data in the classification model; The nonlinear distribution of the target data in the classification model is obtained by introducing a radial basis function kernel function, specifically as follows: ; in, For the first The feature vector of the first core sample and the first The radial basis kernel function values ​​between the eigenvectors of each core sample For the first The target data feature vector corresponding to each core sample For the first The target data feature vector corresponding to each core sample These are the key parameters of the kernel function. For target data feature vectors , Euclidean distance, , The sample number. ; S422. Based on the nonlinear distribution of the target data in the classification model, construct and solve the dual problem to determine the support feature vector of the target data. The target data supporting feature vector is the decisive target data feature vector selected from multiple target data feature vectors, specifically in the following manner: ; in, Indicates the Lagrange multipliers Find the maximum value of the objective function. For the first Lagrange multipliers for each core sample , These represent core samples. , Shale gas reservoir lithofacies type tags; The dual problem is solved by a quadratic programming solver to obtain the optimal Lagrange multiplier vector. The training sample corresponding to the optimal Lagrange multiplier vector being greater than 0 is the target data support feature vector.

[0011] In some feasible implementation methods, the specific method for calculating the optimal offset is as follows: ; in, Let be the index of any support vector, and let the optimal Lagrange multiplier of that support vector be between 0 and the penalty parameter. between, For the first Each support vector corresponds to a shale gas reservoir lithofacies type label. For the first The lithofacies type label corresponding to each support vector. For the first The optimal Lagrange multiplier vector corresponding to each support vector. For the first The support vector feature vector and the first support vector feature vector The radial basis kernel function values ​​between the eigenvectors of the support vectors; Based on the position of the elemental logging data transmitted in real time from the newly drilled formation relative to the translated baseline in the high-dimensional characteristic space, the corresponding lithofacies can be directly determined.

[0012] A second aspect of the present invention provides a shale gas reservoir drilling facies identification system, employing the shale gas reservoir drilling facies identification method described in any one of the above aspects, wherein the identification system further includes: The data acquisition module is used to acquire core sample data and elemental logging data of shale gas reservoirs; The lithofacies classification module is used to classify the lithofacies of the shale gas reservoir according to the classification criteria based on the core sample data, and obtain core lithofacies data. An element selection module is used to select target data that are associated with shale gas reservoirs from the dimensionless element logging data. A classification model building module, which is used to build a model for classifying lithofacies; The lithofacies identification module takes the logging data of the elements to be identified as input into the lithofacies identification module and outputs the lithofacies identification results.

[0013] A third aspect of the present invention also provides a shale gas reservoir drilling facies identification device, the identification device comprising: a processor, a memory, and a data acquisition device; The data acquisition device is connected to the processor and is used to acquire relevant data on the lithofacies of shale gas reservoirs and transmit the acquired data to the processor. The memory is connected to the processor and is used to store intermediate data generated during the processor's data processing as well as the final lithofacies identification results. The processor is used to execute a shale gas reservoir drilling facies identification method as described in any one of the first aspects.

[0014] The beneficial effects of this invention are as follows: This application classifies lithofacies by acquiring core sample data and elemental logging data from shale gas reservoirs, and filters target data related to rocks from the elemental logging data. Since elemental logging is carried out simultaneously with drilling, the lithofacies of the target shale gas reservoir is identified during drilling through a lithofacies identification module. This solves the technical problem of low timeliness of lithofacies identification in the prior art and improves the efficiency and accuracy of lithofacies identification. Attached Figure Description

[0015] Figure 1 This is a schematic flowchart of a shale gas reservoir drilling facies identification method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the three-terminal element projection points provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of a three-terminal element projection point established based on the lithofacies of the Long-1 sub-section of a certain well area, provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of the grey relational degree of logging data for each element provided in the embodiments of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0017] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the meaning of "and / or" throughout the text includes three parallel solutions; for example, "A and / or B" includes solution A, solution B, or a solution where both A and B are satisfied simultaneously. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0018] Reference Figures 1 to 4 As shown, in order to solve the shortcomings of existing technologies where it is difficult to acquire lithofacies data in real time during drilling and the accuracy of the acquired lithofacies data is low, the first aspect of this application provides a method for identifying lithofacies in shale gas reservoirs during drilling, including the following steps; S1. Obtain core sample data and elemental logging data of shale gas reservoirs, and perform dimensionless processing on the elemental logging data; It should be understood that the core samples are several cylindrical rock samples obtained from underground at different depths below the surface through drilling.

[0019] The core sample data is based on a database of the target shale gas reservoir, obtained from historical data or experimental analysis. Specifically, the total organic carbon (TOC) data of the core samples is used to characterize the abundance of organic carbon in various mineral components within the core, and includes X-ray diffraction data and total organic carbon (TOC) data. More specifically, the X-ray diffraction data of the core samples represents the content of various mineral components in the core, such as the relative content of felsic minerals, carbonate minerals, and clay minerals.

[0020] The elemental logging data can be acquired in real time during the drilling process. The elemental logging data includes the content data of various chemical elements in the rock formation and elemental ratio parameters, such as the concentration values ​​of elements like silicon (Si) and iron (Fe), as well as ratio data such as strontium-barium ratio and nickel-cobalt ratio.

[0021] In this embodiment, the purpose of dimensionless processing of the elemental logging data is to eliminate the differences in dimensions and orders of magnitude between different elements, facilitating subsequent processing steps. The specific method for dimensionless processing of the elemental logging data is as follows: ; in, The data consists of elemental logging data after dimensionless processing. For elemental logging data, This represents the maximum content of the same element across all samples. This represents the minimum content of the same element in all samples.

[0022] S2. Based on the core sample data, establish a shale gas reservoir lithofacies classification standard to classify shale gas reservoir lithofacies types; In this embodiment, the establishment of shale gas reservoir lithofacies classification criteria involves setting an organic carbon content threshold and determining different lithologies to classify shale gas reservoir lithofacies. Specifically, this includes the following steps: S21. Based on the organic carbon content data of core samples, set an organic carbon content threshold, and classify the core samples into three organic matter lithofacies according to the organic matter abundance based on the organic carbon content threshold. In this embodiment, based on the organic carbon content data of the core samples, 2% and 3% are used as the organic matter abundance classification boundaries, and an organic carbon content threshold is set. Specifically, when the organic carbon content in the core sample is greater than 3%, it is classified as an organic-rich rock facies; when the organic carbon content in the core sample is in the range of 2% to 3%, it is classified as a medium-organic rock facies; and when the organic carbon content in the core sample is less than 2%, it is classified as a poor-organic rock facies. S22. Based on X-ray diffraction data of core samples, shale gas reservoirs are classified into 10 different types of shale gas reservoir lithology. In this embodiment, based on the different contents of felsic minerals, carbonate minerals, and clay minerals in the core samples, shale gas reservoirs are classified into 10 different shale gas reservoir lithologies. These 10 different shale gas reservoir lithologies include high-siliceous shale, high-calcareous shale, high-argillaceous shale, medium-siliceous shale, medium-calcareous shale, medium-argillaceous shale, siliceous and calcareous mixed shale, calcareous and argillaceous mixed shale, siliceous and argillaceous mixed shale, and impurity mixed shale. This is because felsic minerals (such as quartz and feldspar), carbonate minerals (such as calcite and dolomite), and clay minerals (such as illite and montmorillonite) are key components controlling the physical properties and gas-bearing capacity of shale reservoirs. Their relative contents directly affect the brittleness, pore structure, and organic matter occurrence state of the rock. Therefore, the contents of these three minerals can directly reflect the shale gas reservoir lithology.

[0023] S23. Draw a three-terminal element diagram based on 10 different types of shale gas reservoir lithology; In this embodiment, a three-terminal graph is drawn using the contents of felsic minerals, carbonate minerals, and clay minerals as the three endpoints, such as... Figure 2As shown, the lithological types are defined at 40% and 60% boundaries. Zone I is defined as the core sample containing more than 60% felsic minerals and less than 40% clay minerals. Other lithologies are similarly classified. Specifically, I represents high-siliceous shale, II represents high-calcareous shale, III represents high-argillaceous shale, IV represents medium-siliceous shale, V represents medium-calcareous shale, VI represents medium-argillaceous shale, VII represents mixed silica and calcareous shale, VIII represents mixed calcareous and argillaceous shale, IX represents mixed silica and argillaceous shale, and X represents mixed shale. This is because the three-end-member plot can visually reflect the relative proportions of felsic, carbonate, and clay mineral components, facilitating lithological classification and allowing for rapid determination of the lithological category of each core sample.

[0024] S24. Couple the organic matter abundance classification results with lithological types to establish a shale gas reservoir lithological classification standard and classify shale gas reservoir lithological types. Specifically, by combining the 10 different shale gas reservoir lithologies with the organic matter abundance levels of the 3 types of core samples, 30 comprehensive facies categories of shale gas reservoirs can be obtained. The 30 comprehensive facies categories of shale gas reservoirs are shown in Table 1 below. Table 1

[0025] For example, a lithofacies model was established based on the Longyi-1 sub-member of a certain well area, such as... Figure 3 The three-terminal element plotting diagram shown depicts several core samples. The results indicate that the lithology of the Long-1 sub-member in a certain well area includes three types: high-siliceous shale, medium-siliceous shale, and mixed shale. Based on the corresponding core organic carbon content data, the lithofacies of the Long-1 sub-member in the certain well area are finally determined to be eight types: organic-rich high-siliceous shale facies, organic-rich medium-siliceous shale facies, medium-organic high-siliceous shale facies, medium-organic medium-siliceous shale facies, medium-organic mixed shale facies, organic-poor high-siliceous shale facies, organic-poor medium-siliceous shale facies, and organic-poor mixed shale facies.

[0026] S3. Based on the dimensionless element logging data, target data is filtered through grey relational analysis. It is important to understand that the target data refers to several representative dimensionless elemental logging data selected from all dimensionless processed elemental logging data to characterize the lithofacies of shale gas reservoirs. This is because elemental logging data contains numerous elemental data points. Some elemental data, such as the strontium-barium ratio, nickel-cobalt ratio, and chromium-copper ratio, are closely related to mineral composition. Their changes effectively reflect the relative proportions of felsic minerals, carbonates, and clay minerals. These elemental logging data have a strong correlation with lithofacies, while other elemental logging data have low or no correlation with lithofacies. Therefore, target data can be screened from the dimensionless processed elemental logging data using grey relational analysis, specifically including the following steps: S31. Based on the dimensionless element logging data, obtain the gray correlation coefficient between the core lithofacies and the element logging data. In this embodiment, the grey relational coefficient can be calculated using elemental logging data and core lithofacies, specifically as follows: The core facies sorted by depth after classification are used as a reference sequence: ; in, This is a reference sequence of core facies. For the first The core lithofacies values ​​corresponding to each sample point The index of the data in the sequence. The sequence length is the total number of data points. ; Elemental logging data of the same depth, after being processed without dimension, are sorted by depth as a comparison sequence: ; in, For the first The element logging comparison sequence corresponding to each element. For the first The element in the first The element values ​​at each sample point are recorded. For element type serial number, The number of element types; Calculate the grey relational coefficients between the reference sequence and the comparison sequence: ; in, For the first The element in the first Grey relational coefficient at each sample point This is the two-level minimum absolute difference between the reference sequence corresponding to all elements and all sample points and the sequence of the object to be evaluated. This is the two-level maximum absolute difference between the reference sequence corresponding to all elements and all sample points and the sequence of the object to be evaluated. The resolution coefficient is used to improve the significance of differences between grey relational coefficients. Based on database, historical data, or experimental analysis, the resolution coefficient is set to 0.5.

[0027] S32. Based on the grey relational coefficient, calculate the grey relational degree between elemental logging data and the overall grey relational degree of core lithofacies. The specific method is as follows: ; in, To compare the grey relational degree between the sequence and the reference sequence, if the calculated The higher the value, the higher the correlation between the elemental logging data and the core lithofacies.

[0028] S33. Set a gray relational degree selection threshold, and select target data from several dimensionless element logging data based on the selection threshold; In this embodiment, the selection threshold is set based on the geological characteristics and data quality of different well areas. Based on the set selection threshold, several dimensionless logging data elements with a value greater than the selection threshold are selected as target data from the overall gray relational analysis. For example, as shown... Figure 4 As shown, by calculating the grey relational degree of each element's logging data, logging data of elements with a value greater than 0.5 are set as target data. Thus, logging data of eight elements, namely strontium-barium ratio, nickel-cobalt ratio, chromium-copper ratio, magnesium, silicon, manganese-zinc ratio, potassium-sodium ratio, and titanium, are selected as target data.

[0029] S4. Based on the core facies type and target data, obtain drilling facies identification information of the target shale gas reservoir; It should be understood that the target shale gas reservoir is a shale gas reservoir that requires drilling operations according to production requirements.

[0030] In this embodiment, the acquisition of drilling facies identification information of the target shale gas reservoir can be achieved by establishing a training dataset of core facies types and target data, and then establishing a classification model based on the training dataset to classify the facies types of the target shale gas reservoir, and finally identifying the classification results and outputting the identification information.

[0031] The process of obtaining drilling-while-drilling facies identification information for the target shale gas reservoir based on core facies type and target data includes the following steps: S41. Construct a core sample training dataset based on core lithofacies type and target data, specifically as follows: ; in, For core sample training dataset, This refers to the number of core samples. For the first The target data feature vector corresponding to each core sample For the first The lithological type label of the shale gas reservoir corresponding to each core sample. This indicates that the core samples are arranged by serial number. From 1 to A set of samples arranged sequentially; S42. Based on the core sample training dataset, construct a classification model to classify the lithofacies of the core samples; In some implementations, a support vector machine (SVM) classification model is constructed based on the training data of each core sample in the core sample training dataset. Specifically: ; in, For the weight vector, This is the offset. The set of relaxation variables corresponding to all core samples. This represents finding the minimum value over the weight vector, the offset, and the set of relaxation variables corresponding to all core samples. This is the penalty parameter, used to balance maximizing the classification margin and minimizing the training error. For the first Relaxation variables corresponding to each core sample Indicates all The relaxation variables of each core sample are summed. To mark the constraints, For weight transpose, This is a high-dimensional mapping function; the relaxation variables and constraints are set through database, historical data, or experimental analysis. The reason for this is that there may be misalignment between the core sampling depth and the elemental logging sampling depth of different core samples, leading to inaccurate data calibration and errors. Therefore, by setting relaxation variables and penalty parameters, frequent misjudgments by the model can be avoided, thus improving model accuracy.

[0032] In some implementations, since the differences between different core sample data are similar, while the differences between elemental logging data are large, in order to further improve the accuracy of the classification model, the construction of the classification model based on the core sample training dataset also includes: S421. Obtain the nonlinear distribution of the target data in the classification model; The purpose of obtaining the nonlinear distribution of target data in the classification model is to more intuitively reflect the characteristics and distribution of different lithofacies types, thereby optimizing the classification model and making its lithofacies classification more accurate.

[0033] The nonlinear distribution of the target data in the classification model can be obtained by introducing a radial basis function kernel function, specifically as follows: ; in, For the first The feature vector of the first core sample and the first The radial basis kernel function values ​​between the eigenvectors of each core sample For the first The target data feature vector corresponding to each core sample For the first The target data feature vector corresponding to each core sample These are the key parameters of the kernel function. For target data feature vectors , Euclidean distance, The sample number. In this embodiment, a radial basis function kernel is introduced and key parameters are determined. The range of values ​​allows for a more intuitive understanding of the nonlinear distribution of target data in the classification model.

[0034] S422. Based on the nonlinear distribution of the target data in the classification model, construct and solve the dual problem to determine the support feature vector of the target data. The target data support feature vector is the target data feature vector selected from multiple target data feature vectors that plays a decisive role; that is, the target data feature vector that determines the boundary position in the classification model. Specifically, it is implemented as follows: ; in, Indicates the Lagrange multipliers Find the maximum value of the objective function. For the first Lagrange multipliers for each core sample , These represent core samples. , Tag of lithological type for shale gas reservoirs.

[0035] In this embodiment, the dual problem described above is solved using a quadratic programming solver. This quadratic programming solver includes, but is not limited to, interior-point methods, effective set methods, and dual methods—solvers capable of solving quadratic programming problems of arbitrary scales—to obtain the optimal Lagrange multiplier vector. The training samples corresponding to the optimal Lagrange multiplier vector being greater than 0 are the target data support feature vectors. Obtaining the target data support feature vectors reduces the frequency of model calls, lowers the computational load, and makes the classification model faster in identifying lithofacies.

[0036] S43. Based on the classification model, obtain drilling facies identification information of the target shale gas reservoir; In this embodiment, by constructing a decision function, the dimensionless elemental logging data of the target shale gas reservoir is input into the classification model, and the lithofacies identification information of the target shale gas reservoir is output through the decision function.

[0037] The decision function is: ; in, This is a feature vector for identifying the lithofacies information of the target shale gas reservoir. This is a function for determining the sign of a sign. To support the set of feature vectors for the target data, This indicates that the summation is performed only on samples within the support vector set. For optimal Lagrange multipliers, This is the optimal offset; if If the value is greater than 0, then output the identification information of a certain type of lithofacies; if... If the value is less than 0, other types of rock facies identification information will be output.

[0038] The reason for this is that elemental logging data (such as silicon content, magnesium content, and ratios of various elements) are absolute quantitative indicators with clear physical meaning and consistently positive values. In the multidimensional feature space, all lithofacies samples are not centered around the origin, but rather exist in a numerical range far from the origin. If no intercept term is set for the decision boundary, the resulting boundary will deviate significantly from the actual boundary areas of different shale lithofacies (such as "organic-rich, high-siliceous shale facies" and "organic-poor, medium-siliceous shale facies"), making accurate lithofacies division impossible. Therefore, the optimal offset can be calculated. To obtain an optimal boundary baseline, more specifically, the method for calculating the optimal offset is as follows: ; in, Let be the index of any support vector, and let the optimal Lagrange multiplier of that support vector be between 0 and the penalty parameter. between, For the first Each support vector corresponds to a shale gas reservoir lithofacies type label. For the first The lithofacies type label corresponding to each support vector. For the first The optimal Lagrange multiplier vector corresponding to each support vector. For the first The support vector feature vector and the first support vector feature vector The radial basis kernel function values ​​between the eigenvectors of the support vectors; In this embodiment, for the elemental logging data transmitted in real time from the newly drilled formation, the corresponding lithofacies can be directly determined based on its position (positive or negative side region) relative to the translated baseline in the high-dimensional feature space. This method has high identification accuracy and enables real-time lithofacies classification.

[0039] The second aspect of this application also provides a shale gas reservoir drilling facies identification system in its embodiments, employing the shale gas reservoir drilling facies identification method described in any one of the first aspects above. The identification system further includes: a data acquisition module, a facies classification module, an element selection module, a classification model construction module, and a facies identification module. Specifically, the data acquisition module is used to acquire core sample data and element logging data of the target shale gas reservoir; the facies classification module is used to classify the shale gas reservoir facies according to classification standards based on the core sample data to obtain core facies data; the element selection module is used to select target data that are associated with the shale gas reservoir from the dimensionless element logging data; the classification model construction module is used to establish a model for classifying the facies; since element logging is performed simultaneously with drilling, the facies identification module can input the element logging data of the shale gas well to be identified and then output the facies identification result.

[0040] A third aspect of this invention provides a shale gas reservoir drilling facies identification device. The device includes a processor, a memory, and a data acquisition device. The data acquisition device is connected to the processor and is used to acquire relevant data about the shale gas reservoir and transmit the acquired data to the processor. The memory is connected to the processor and is used to store intermediate data generated during data processing and the final identification result data. The processor is used to execute a shale gas reservoir drilling facies identification method as described in any one of the first aspects. Specifically, the identification device processes the data acquired by the data acquisition device, including processing shale reservoir core sample data, dimensionless processing of elemental logging data, drawing a three-terminal element plot, classifying facies categories, filtering target data in elemental logging data, and constructing a classification model. Finally, the facies identification result is stored in the memory, and the data in the memory can be retrieved and displayed as needed to identify the facies of the target shale gas reservoir.

[0041] A fourth aspect of the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the shale gas reservoir drilling facies identification method described in any one of the first aspects. The computer-readable storage medium according to embodiments of the present disclosure 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 disclosure are performed. 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).

[0042] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for identifying facies in shale gas reservoirs during drilling, characterized in that, The identification method includes the following steps: S1. Obtain core sample data and elemental logging data of shale gas reservoirs, and perform dimensionless processing on the elemental logging data; The core sample data is based on the database of the target shale gas reservoir, historical data or experimental analysis, and the core sample data includes core sample X-ray diffraction data and core sample organic carbon content data. S2. Based on the core sample data, establish a shale gas reservoir lithofacies classification standard to classify shale gas reservoir lithofacies types; S3. Based on the dimensionless element logging data, target data is filtered through grey relational analysis. S4. Based on the core facies type and target data, obtain drilling facies identification information for the target shale gas reservoir.

2. The method for identifying facies in shale gas reservoirs during drilling according to claim 1, characterized in that, The specific method for performing dimensionless processing on the elemental logging data is as follows: ； in, The data consists of elemental logging data after dimensionless processing. For elemental logging data, This represents the maximum content of the same element across all samples. This represents the minimum content of the same element in all samples.

3. The method for identifying facies in shale gas reservoirs during drilling according to claim 1, characterized in that, S2 specifically includes the following steps: S21. Based on the organic carbon content data of core samples, set an organic carbon content threshold, and classify the core samples into three organic matter lithofacies according to the organic matter abundance based on the organic carbon content threshold. Based on the organic carbon content data of the core samples, 2% and 3% were used as the boundary for organic matter abundance classification, and an organic carbon content threshold was set. When the organic carbon content in the core sample is greater than 3%, it is classified as an organic-rich rock facies; when the organic carbon content in the core sample is in the range of 2% to 3%, it is classified as a medium-organic rock facies; and when the organic carbon content in the core sample is less than 2%, it is classified as an organic-poor rock facies. S22. Based on X-ray diffraction data of core samples, shale gas reservoirs are classified into 10 different types of shale gas reservoir lithology. Based on the different contents of felsic minerals, carbonate minerals, and clay minerals in the core samples, shale gas reservoirs are classified into 10 different shale gas reservoir lithologies. These 10 different shale gas reservoir lithologies include high-siliceous shale, high-calcareous shale, high-muddy shale, medium-siliceous shale, medium-calcareous shale, medium-muddy shale, siliceous and calcareous mixed shale, calcareous and muddy mixed shale, siliceous and muddy mixed shale, and impurity mixed shale. S23. Draw a three-terminal element diagram based on 10 different types of shale gas reservoir lithology; A three-endpoint diagram was drawn using the contents of felsic minerals, carbonate minerals, and clay minerals as the three endpoints. S24. Couple the organic matter abundance classification results with lithological types to establish a shale gas reservoir lithological classification standard and classify shale gas reservoir lithological types. By combining 10 different shale gas reservoir lithologies with 3 organic matter facies, 30 comprehensive facies categories of shale gas reservoirs were obtained.

4. The method for identifying facies in shale gas reservoirs during drilling according to claim 1, characterized in that, The process of filtering target data through grey relational analysis specifically includes the following steps: S31. Based on the dimensionless elemental logging data, obtain the grey relational coefficient between the core lithofacies and the elemental logging data. The specific method is as follows: The core facies sorted by depth after classification are used as a reference sequence: ； in, This is a reference sequence of core facies. For the first The core lithofacies values ​​corresponding to each sample point The index of the data in the sequence. The sequence length is the total number of data points. ; Elemental logging data of the same depth, after being processed without dimension, are sorted by depth as a comparison sequence: ； in, For the first The dimensionless element logging comparison sequence corresponding to each element. For the first The element in the first Dimensionless element logging values ​​at each sample point For element type serial number, The number of element types; Calculate the grey relational coefficients between the reference sequence and the comparison sequence: ； in, For the first The element in the first Grey relational coefficient at each sample point This is the two-level minimum absolute difference between the reference sequence corresponding to all elements and all sample points and the sequence of the object to be evaluated. This is the two-level maximum absolute difference between the reference sequence corresponding to all elements and all sample points and the sequence of the object to be evaluated. The resolution coefficient is set to 0.5 based on database, historical data, or experimental analysis. S32. Based on the grey relational coefficient, calculate the grey relational degree between element logging data and the overall grey relational degree between core lithofacies; The specific method is as follows: ； in, To compare the grey relational degree between the sequence and the reference sequence, if the calculated The higher the value, the higher the correlation between the element logging data and the core lithofacies. S33. Set a gray relational degree selection threshold, and select target data from several dimensionless element logging data based on the selection threshold; The selection threshold is set according to the geological characteristics and data quality of different well areas, and based on the set selection threshold, several types of logging data with a value greater than the selection threshold are selected as target data from the overall gray relational degree.

5. The method for identifying facies in shale gas reservoirs during drilling according to claim 1, characterized in that, S4 includes the following steps: S41. Construct a core sample training dataset based on core lithofacies type and target data, specifically as follows: ； in, For core sample training dataset, This refers to the number of core samples. For the first The target data feature vector corresponding to each core sample For the first The lithological type label of the shale gas reservoir corresponding to each core sample. This indicates that the core samples are arranged by serial number. From 1 to A set of samples arranged sequentially; S42. Based on the core sample training dataset, construct a classification model to classify the lithofacies of the core samples; Based on the training data of each core sample in the core sample training dataset, a support vector machine (SVM) classification model is constructed. The specific method is as follows: ； in, For the weight vector, This is the offset. The set of relaxation variables corresponding to all core samples. This represents finding the minimum value over the weight vector, the offset, and the set of relaxation variables corresponding to all core samples. This is the penalty parameter, used to balance maximizing the classification margin and minimizing the training error. For the first Relaxation variables corresponding to each core sample Indicates all The relaxation variables of each core sample are summed. To mark the constraints, For weight transpose, It is a high-dimensional mapping function; where slack variables and constraints are set through databases, historical data, or experimental analysis; S43. Based on the classification model, obtain drilling facies identification information of the target shale gas reservoir; By constructing a decision function, the dimensionless elemental logging data of the target shale gas reservoir is input into the classification model, and the lithofacies identification information of the target shale gas reservoir is output through the decision function. The decision function is: ； in, This is a feature vector for identifying the lithofacies information of the target shale gas reservoir. This is a function for determining the sign of a sign. To support the set of feature vectors for the target data, This indicates that the summation is performed only on samples within the support vector set. For optimal Lagrange multipliers, This is the optimal offset; if If the value is greater than 0, then output the identification information of a certain type of lithofacies; if... If the value is less than 0, other types of rock facies identification information will be output.

6. The method for identifying facies in shale gas reservoirs during drilling according to claim 5, characterized in that, The classification model constructed based on the core sample training dataset also includes: S421. Obtain the nonlinear distribution of the target data in the classification model; The nonlinear distribution of the target data in the classification model is obtained by introducing a radial basis function kernel function, specifically as follows: ； in, For the first The feature vector of the first core sample and the first The radial basis kernel function values ​​between the eigenvectors of each core sample For the first The target data feature vector corresponding to each core sample For the first The target data feature vector corresponding to each core sample These are the key parameters of the kernel function. For target data feature vectors , Euclidean distance, , The sample number. ; S422. Based on the nonlinear distribution of the target data in the classification model, construct and solve the dual problem to determine the support feature vector of the target data. The target data supporting feature vector is the decisive target data feature vector selected from multiple target data feature vectors, specifically in the following manner: ； in, Indicates the Lagrange multipliers Find the maximum value of the objective function. For the first Lagrange multipliers for each core sample , These represent core samples. , Shale gas reservoir lithofacies type tags; The dual problem is solved by a quadratic programming solver to obtain the optimal Lagrange multiplier vector. The training sample corresponding to the optimal Lagrange multiplier vector being greater than 0 is the target data support feature vector.

7. The method for identifying facies in shale gas reservoirs during drilling according to claim 5, characterized in that, The specific method for calculating the optimal offset is as follows: ； in, Let be the index of any support vector, and let the optimal Lagrange multiplier of that support vector be between 0 and the penalty parameter. between, For the first Each support vector corresponds to a shale gas reservoir lithofacies type label. For the first The lithofacies type label corresponding to each support vector. For the first The optimal Lagrange multiplier vector corresponding to each support vector. For the first The support vector feature vector and the first support vector feature vector The radial basis kernel function values ​​between the eigenvectors of the support vectors; Based on the position of the elemental logging data transmitted in real time from the newly drilled formation relative to the translated baseline in the high-dimensional characteristic space, the corresponding lithofacies can be directly determined.

8. A shale gas reservoir drilling facies identification system, characterized in that, The method for identifying shale gas reservoir facies during drilling, as described in any one of claims 1 to 7, further includes: The data acquisition module is used to acquire core sample data and elemental logging data of shale gas reservoirs; The lithofacies classification module is used to classify the lithofacies of the shale gas reservoir according to the classification criteria based on the core sample data, and obtain core lithofacies data. An element selection module is used to select target data that are associated with shale gas reservoirs from the dimensionless element logging data. A classification model building module, which is used to build a model for classifying lithofacies; The lithofacies identification module takes the logging data of the elements to be identified as input into the lithofacies identification module and outputs the lithofacies identification results.

9. A shale gas reservoir drilling facies identification device, characterized in that, The identification device includes: a processor, a memory, and a data acquisition device; The data acquisition device is connected to the processor and is used to acquire relevant data on the lithofacies of shale gas reservoirs and transmit the acquired data to the processor. The memory is connected to the processor and is used to store intermediate data generated during the processor's data processing as well as the final lithofacies identification results. The processor is used to execute a shale gas reservoir drilling facies identification method as described in any one of claims 1 to 7.

Images