Marine facies shale lithofacies logging fine identification method and device

Abstract

The invention discloses a marine shale lithofacies logging fine identification method and device, and belongs to the technical field of oil-gas exploration, and the method comprises the steps: determining a marine shale lithofacies division scheme; determining a target layer needing to be subjected to marine facies shale lithofacies division, and obtaining logging data of the target layer; the mineral content and the organic matter content of the rock core at each depth determined at equal intervals on the target layer are obtained; aiming at each logging curve in the target layer logging data, screening based on correlation analysis; constructing a sample data set; optimizing the RBF neural network by utilizing a clustering algorithm, taking the sample data as input, taking the mineral and organic matter contents at the depth corresponding to the sample data as output, and respectively training to obtain a mineral content prediction model and an organic matter content prediction model; and performing lithofacies fine division according to output data of the prediction model. According to the method, under the condition of less core data, non-linear high-precision prediction of the mineral content and the organic matter content is achieved, and fine identification of marine shale lithofacies logging is achieved.

Description

Technical Field

[0001] The invention relates to a marine shale lithofacies logging fine identification method, and in particular to a marine shale lithofacies logging fine identification method and device. Background Art

[0002] In recent years, due to the rapid development of unconventional oil and gas exploration and development, shale reservoir evaluation has received widespread attention worldwide. The prediction of high-quality shale reservoirs is the basis for efficient exploration and development of shale gas. Shale reservoirs are highly heterogeneous, and different lithofacies have different development potentials. The hydrocarbon generation capacity and reservoir performance of shale lithofacies are affected by factors such as organic matter abundance and mineral content. Therefore, lithofacies are closely related to oil and gas content. The division of lithofacies types and the identification of favorable shale lithofacies sections are crucial for shale reservoir prediction. Therefore, an accurate and reliable shale lithofacies fine identification method plays a very important role in the exploration and development of unconventional oil and gas reservoirs.

[0003] At present, marine shale lithofacies are mainly divided by combining organic matter and mineral content. However, drilling coring and testing and analysis are expensive, time-consuming and labor-intensive, and it is difficult to provide a large amount of core data for lithofacies prediction. Therefore, how to use relatively economical and massive geophysical data to complete the effective identification and high-precision prediction of lithofacies is a technical problem that needs to be solved urgently. Summary of the invention

[0004] The present invention aims to solve the above-mentioned problems existing in the prior art for identifying marine shale lithofacies, and proposes a method and device for fine identification of marine shale lithofacies by well logging. The method and device can realize nonlinear high-precision prediction of mineral and organic matter content when there is less core data, and realize fine identification of marine shale lithofacies by well logging.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution of the present invention is as follows:

[0006] A marine shale lithofacies logging fine identification method, characterized by comprising the following steps:

[0007] Step 1: Determine the marine shale lithofacies division scheme of “mineral composition + TOC content”;

[0008] Step 2: determine the target layer for marine shale lithofacies division and obtain the logging data of the target layer; determine multiple depths at equal intervals on the target layer and obtain the mineral content and organic matter content of the core at each depth;

[0009] Step 3: For each logging curve in the target layer logging data, a correlation analysis is performed with the mineral content and the organic matter content to obtain the correlation coefficient between each logging curve and the mineral content and the organic matter content, and the logging curve is selected according to the correlation coefficient;

[0010] Step 4: Select data from target layers of multiple wells and construct a sample data set after normalization.

[0011] Step 5: Optimize the RBF neural network using the K-means clustering algorithm, take the sample data as input, and the mineral and organic matter contents at the depth corresponding to the sample data as output, and train the prediction models of mineral content and organic matter content respectively;

[0012] Step 6: Perform detailed lithology division based on the output data of the prediction model.

[0013] Furthermore, the marine shale lithofacies division scheme determined in step 1 is:

[0014] Marine shale lithofacies are divided into two categories based on the TOC content of 2%: organic-rich shale with a TOC content ≥ 2% and organic-poor shale with a TOC content < 2%;

[0015] According to the main mineral composition of shale, the contents of carbonate minerals, clay minerals and siliceous minerals are used as the three end members of the triangle chart; calcite and dolomite represent carbonate minerals; feldspar and quartz represent siliceous minerals; montmorillonite, kaolinite, illite-montmorillonite mixed layer, chlorite-montmorillonite mixed layer, illite and chlorite represent clay minerals;

[0016] The shale phases with a content exceeding 50% of the dividing line are named calcareous shale phase, siliceous shale phase, and clay shale phase, and the shale phases with a mineral content not exceeding 50% are named mixed shale phase.

[0017] Furthermore, in step 2, the logging data of the target layer includes a variety of logging curves including density, natural gamma, porosity, shear wave time difference, potassium element, uranium element, uranium-free gamma, and water saturation.

[0018] Furthermore, in step 2, the mineral content of the core is obtained by whole-rock analysis of the core using X-ray diffraction, and the total organic carbon content of the core is obtained by geochemical analysis.

[0019] Furthermore, in step three, SPSS statistical analysis software was used to conduct correlation analysis.

[0020] Furthermore, in step three, by presetting a threshold, the logging curves with correlation coefficients greater than the threshold are retained to participate in subsequent steps.

[0021] Furthermore, in step 4, the sample data set constructed is l j = {l 1j , l 2j ,…,l ij ,…,l Nj}; where l ijIt represents the value of the i-th logging curve at the j-th depth, i=1~N, j=1~M, N is the number of logging curves, and M is the number of multiple depths determined at equal intervals on the target layer.

[0022] The present invention also provides a marine shale lithofacies logging fine identification device, comprising:

[0023] Shale lithofacies division module, used to determine the marine shale lithofacies division scheme of "mineral composition + TOC content";

[0024] The data acquisition module is used to determine the target layer for marine shale lithofacies division and obtain the logging data of the target layer; determine multiple depths at equal intervals on the target layer and obtain the mineral content and organic matter content of the core at each depth;

[0025] The logging curve screening module is used to perform correlation analysis on each logging curve in the target layer logging data with the mineral content and organic matter content, obtain the correlation coefficient between each logging curve and the mineral content and organic matter content, and select the logging curve according to the correlation coefficient;

[0026] The sample data construction module is used to select data from target layers of multiple wells and construct a sample data set after normalization processing;

[0027] The model training module is used to optimize the RBF neural network using the K-means clustering algorithm, taking sample data as input and the mineral and organic matter contents at the corresponding depth of the sample data as output, and respectively training to obtain prediction models for mineral content and organic matter content.

[0028] The present invention also provides a marine shale lithofacies logging fine identification device, comprising a processor and a memory for storing processor executable instructions, wherein the instructions, when executed by the processor, implement the steps of the above-mentioned marine shale lithofacies logging fine identification method.

[0029] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned marine shale lithofacies logging fine identification method are implemented.

[0030] In summary, the present invention has the following advantages:

[0031] 1. Based on the perspective of shale mineral composition and organic matter content, the present invention establishes a nonlinear mapping relationship between logging curves and mineral content based on the K-means-RBF neural network for the first time, thus realizing low-cost and high-precision logging identification of heterogeneous marine shale lithofacies;

[0032] 2. The present invention optimizes the RBF neural network using the K-means clustering algorithm, solves the problem that the center and width of the radial basis function of the RBF neural network are difficult to determine, and greatly improves the stability and reliability of the RBF neural network model. The combined K-means-RBF neural network can train a complex nonlinear relationship network model between mineral and organic matter content and other multiple parameters, improves prediction accuracy, realizes high-precision prediction of marine shale mineral and organic matter content, and completes fine division and identification of shale lithofacies;

[0033] 3. The present invention has high prediction accuracy and obvious recognition effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a flow chart of the implementation method of the present invention;

[0035] Figure 2 It is a diagram of the lithofacies division pattern of marine shale;

[0036] Figure 3 This is a comparison chart of the prediction results of the clay mineral content training set and the test set;

[0037] Figure 4 This is a comparison chart of the prediction results of the calcium mineral content training set and the test set;

[0038] Figure 5 This is a comparison chart of the prediction results of the training set and the test set for siliceous mineral content;

[0039] Figure 6 This is a comparison chart of the prediction results of the organic matter content training set and the test set;

[0040] Figure 7 This is a comparison chart of the shale lithofacies identification effect of the K-means-RBF neural network model. DETAILED DESCRIPTION

[0041] In order to explain the present invention more clearly, the present invention is further described below in conjunction with preferred embodiments and drawings. It should be understood by those skilled in the art that the content described below is illustrative rather than restrictive, and should not be used to limit the scope of protection of the present invention.

[0042] Example 1

[0043] This embodiment provides a marine shale lithofacies logging fine identification method based on K-means-RBF neural network, such as Figure 1 The following steps are shown:

[0044] Step 1: Determine the marine shale lithofacies division scheme of "mineral composition + TOC content" through investigation and research.

[0045] like Figure 2 As shown, the present invention determines the marine shale lithofacies division scheme of "mineral composition + TOC content", specifically, it is divided into two categories with TOC content of 2% as the boundary (TOC content ≥ 2% is organic-rich shale, TOC content < 2% is organic-poor shale), and then according to the main mineral composition of shale, the contents of carbonate minerals, clay minerals and siliceous minerals are used as the three end members of the triangle plate: calcite and dolomite represent carbonate minerals; feldspar and quartz represent siliceous minerals; montmorillonite, kaolinite, illite-montmorillonite mixed layer, chlorite-montmorillonite mixed layer, illite and chlorite represent clay minerals; those with contents exceeding 50% of the dividing line are named calcareous shale phase, siliceous shale phase and clayey shale phase respectively, and the shale lithofacies with mineral contents not exceeding 50% are named mixed shale phase.

[0046] Step 2: determine the target layer section of the formation for the mineral content and organic matter content to be measured, and obtain the logging data of the target layer, which includes multiple logging curves, such as density, natural gamma, porosity, shear wave time difference, potassium, uranium, uranium-free gamma, and water saturation;

[0047] Determine M depths at equal intervals on the target layer, and obtain the mineral and organic content of the core at each depth. The mineral content of the core is obtained by whole-rock X-ray diffraction analysis of shale core data, and the total organic carbon content is obtained by geochemical analysis.

[0048] Assuming that the target layer is located at a depth of 3000-3100 meters underground, the total thickness of the target layer is 100 meters. We sample at equal intervals of 1 meter, so there are 100 sampling points at depth.

[0049] Step 3: Perform correlation analysis on each well logging curve with mineral and organic matter content to obtain the correlation coefficient between each well logging curve and mineral content and organic matter content. A threshold is preset based on experience, and well logging curves with correlation coefficients greater than the threshold are retained. Suppose there are N types of well logging curves, which are marked as 1st to Nth types respectively.

[0050] In this step, based on experience, the threshold can be set above 0.4 or 0.5. There are dozens to hundreds of logging curves in the logging data. It is a large workload to conduct correlation analysis between each logging curve and the mineral and organic matter content. In order to simplify the calculation amount, a batch of logging curves can be preliminarily screened based on experience, and then the correlation analysis between the preliminarily screened logging curves and the mineral and organic matter content is performed, and the logging curves with correlation coefficients greater than the threshold are retained.

[0051] In this embodiment, the following logging curves are analyzed and found to be highly correlated with the mineral or organic matter content, including: DEN (density), GR (natural gamma), POR (porosity), DTS (shear time difference), K (potassium element), U (uranium element), KTH (uranium-free gamma), SW (water saturation), and the correlation is shown in Table 1.

[0052] Table 1 Correlation analysis between logging curves and mineral content and organic matter

[0053]

[0054] The preferred logging curves are selected based on correlation analysis. The preferred logging curves for clay minerals include: DEN, GR, DTS, KTH, U; the preferred logging curves for calcareous minerals include: DEN, DTS, K, KTH, U; and the preferred logging curves for organic matter include: DEN, GR, K, POR, U, SW. Since the correlation of siliceous minerals is low, it is not suitable to directly establish a prediction model. Studies have shown that in sedimentary rocks, the main mineral components include siliceous minerals, carbonate minerals and clay minerals, and the three components basically occupy the entire rock skeleton. Therefore, it can be considered that the sum of the contents of clay minerals, carbonate minerals and siliceous minerals is 100%, that is, the content of siliceous minerals can be calculated using the following equation:

[0055] Siliceous=100%-Caly-Carbonate;

[0056] In the formula: Siliceous is the content of siliceous minerals; Clay is the content of clay minerals; Carbonate is the content of carbonate minerals.

[0057] Step 4: Construct sample data;

[0058] The data of target layers of multiple wells are selected to construct sample data. After normalization, all sample data are combined into a sample data set. The constructed data set is L j ={L 1j , L 2j , …, L ij , …, L Nj}; where L ij It represents the value of the i-th logging curve at the j-th depth, i = 1 to N, j = 1 to M. The mineral composition and organic matter content corresponding to the logging curve are yj and zj respectively.

[0059] For example, the sample data at the first depth L1 = {L 11 , L 21 , L 31 , L 41 , L 51 ,};L 11 -L51 They represent the density, shear wave time difference, potassium, uranium-free gamma, and uranium values ​​of the clay mineral model at this depth. The sample data selection of the calcium and organic matter content models can be obtained in the same way.

[0060] Step 5. Establish a K-means-RBF neural network with the center and width of the radial basis function determined by the K-means clustering algorithm, take the training data as input, and the mineral and organic matter content at the corresponding depth of the training data as output, and train the K-means-RBF neural network mineral and organic matter content prediction models respectively.

[0061] The main purpose of the K-means clustering algorithm in the present invention is to divide data into several categories so that the data in the same category have as high a similarity as possible. The distance between samples is usually used in the algorithm to represent the similarity between them, wherein the smaller the distance is, the higher the similarity between samples is, and vice versa, the smaller the similarity is, the greater the difference is. The RBF neural network consists of an input layer, a hidden layer, and an output layer, and has the significant advantages of optimal fitting performance and global fitting, and is very suitable for fitting nonlinear data.

[0062] The K-means clustering algorithm solves the problem of difficulty in determining the center and width of the radial basis function of the RBF neural network, greatly improving the stability and reliability of the RBF neural network model. The combined K-means-RBF neural network can train a complex nonlinear relationship network model between mineral and organic matter content and multiple other parameters, improve prediction accuracy, achieve high-precision prediction of marine shale mineral and organic matter content, and complete the fine division and identification of shale lithofacies.

[0063] This step specifically includes:

[0064] 80% of the data in the data set lj is used as training data to create a neural network model, and the other 20% of the data is used as test data to verify the reliability of the model;

[0065] Input the training data into the K-means clustering algorithm to determine the center and width of the radial basis function of the RBF neural network;

[0066] The K-means-RBF neural network was established with the center and width of the radial basis function determined by the K-means clustering algorithm. The training data was used as input, and the mineral and organic matter contents at the corresponding depths of the training data were used as output. The K-means-RBF neural network mineral and organic matter content prediction models were obtained by training respectively.

[0067] In order to obtain more sample data, this embodiment selects seven wells in a shale exploration area, collects data from the seven wells to construct sample data, and then constructs all sample data into a data set. Then the sample data in the data set is divided into training data and test data. In actual operation, 80% of the data is used as training data to create a neural network model, and the other 20% of the data is used as test data to verify the reliability of the model. During training, the sample data is used as input, and the mineral and organic content at the corresponding depth of the sample data is output.

[0068] When obtaining training data or test data, the measured core mineral and organic matter contents are used as the measured values, and the corresponding mineral and organic matter contents output in the prediction model are used as the predicted values. The measured values ​​and predicted values ​​of the clay mineral content training set and the test set are compared to obtain Figure 3 , the measured values ​​and predicted values ​​of the calcium mineral content training set and test set are compared to obtain Figure 4 , the measured values ​​and predicted values ​​of organic matter content in the training set and the test set are compared to obtain Figure 5 , by calculating the predicted value of the silica mineral content test set and comparing it with the measured value Figure 6 From the results in the figure, it can be seen that the predicted value is very close to the measured value, with high prediction accuracy. The prediction results of the test data show that the model is stable and reliable, with strong and stable prediction ability.

[0069] According to the predicted data and combined with the lithofacies division scheme, the marine shale lithofacies division is carried out, such as Figure 7 As shown in Table 2, the prediction accuracy of the invention is high, and the accuracy of calcareous shale phase identification is 71.4%, the accuracy of clayey shale phase identification is 89.5%, the accuracy of mixed shale phase identification is 86.6%, the accuracy of siliceous shale phase identification is 83%, and the overall identification accuracy is 85%, with obvious identification effect, as shown in Table 2. In actual production, a K-means-RBF neural network prediction model can be established using several typical wells in a certain area to predict the mineral and organic matter content of other wells in the area, and combined with the lithofacies division scheme, the lithofacies type can be quantitatively identified, thereby improving exploration efficiency and reducing exploration costs.

[0070] Table 2 Statistics of marine shale lithofacies identification accuracy

[0071]

[0072] Example 2

[0073] This embodiment provides a marine shale lithofacies logging fine identification device, comprising:

[0074] The shale lithofacies division module is used to determine the marine shale lithofacies division scheme of "mineral composition + TOC content"; the marine shale lithofacies division scheme of "mineral composition + TOC content" is specifically divided into two categories based on the TOC content of 2% (TOC content ≥ 2% is organic-rich shale, and TOC content < 2% is organic-poor shale), and then according to the main mineral composition of shale, the contents of carbonate minerals, clay minerals and siliceous minerals are used as the three end members of the triangle plate: calcite and dolomite represent carbonate minerals; feldspar and quartz represent siliceous minerals; montmorillonite, kaolinite, illite-montmorillonite mixed layer, chlorite-montmorillonite mixed layer, illite and chlorite represent clay minerals; those with a content exceeding 50% of the dividing line are named calcareous shale phase, siliceous shale phase, and clayey shale phase, and shale phases with mineral contents not exceeding 50% are named mixed shale phase.

[0075] The data acquisition module is used to determine the target layer for marine shale lithofacies division and obtain the logging data of the target layer; determine multiple depths at equal intervals on the target layer and obtain the mineral content and organic matter content of the core at each depth; the logging data of the target layer includes multiple logging curves, such as density, natural gamma, porosity, shear wave time difference, potassium element, uranium element, uranium-free gamma, water saturation, etc.

[0076] The logging curve screening module is used to perform correlation analysis on each logging curve in the target layer logging data with the mineral content and organic matter content, obtain the correlation coefficient between each logging curve and the mineral content and organic matter content, and select the logging curve based on the correlation coefficient. When selecting the logging curve, a threshold can be preset based on experience, and the logging curves with correlation coefficients greater than the threshold are retained.

[0077] The sample data construction module is used to select data from target layers of multiple wells, and construct a sample data set after normalization. The sample data set includes the corresponding mineral composition and organic matter content of all logging curves at each determined depth.

[0078] The model training module is used to optimize the RBF neural network using the K-means clustering algorithm, taking the sample data as input and the mineral and organic matter contents at the corresponding depth of the sample data as output, and respectively training to obtain the prediction models of mineral content and organic matter content;

[0079] The lithofacies division module is used to perform detailed lithofacies division based on the output data of the prediction model.

[0080] The functions implemented by the above-mentioned functional modules may correspond to the various steps of a marine shale lithofacies logging fine identification method in Example 1, and will not be repeated for the undescribed parts.

[0081] Corresponding to the above-mentioned method and device for fine identification of marine shale lithofacies logging, the present invention also provides a marine shale lithofacies logging fine identification device, including a processor and a memory for storing processor executable instructions, and the instructions implement the steps of the above-mentioned method when executed by the processor.

[0082] Preferably, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments that can complete specific functions, and the instruction segments are used to describe the execution process of the computer program in the device.

[0083] The processor can be a central processing unit, or other general-purpose processors, digital signal processors, application-specific integrated circuits, off-the-shelf programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor, or the processor can also be any conventional processor. The processor is the control center of the device, and various parts of the device are connected using various interfaces and lines.

[0084] The memory mainly includes a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function, etc., and the data storage area can store related data, etc. In addition, the memory can be a high-speed random access memory, or a non-volatile memory, such as a plug-in hard disk, a smart memory card, a secure digital card, and a flash memory card, etc., or the memory can also be other volatile solid-state storage devices.

[0085] Corresponding to the above-mentioned marine shale lithofacies logging fine identification method and device, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and the computer program implements the steps of the above-mentioned method when executed by a processor.

[0086] Computer storage media may be tangible media that can contain or store programs for use by or in connection with an instruction execution system, apparatus, or device.

[0087] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any simple modification or equivalent change made to the above embodiment based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A marine shale lithofacies logging fine identification method, characterized in that: The steps include: Step 1: Determine the marine shale lithofacies division scheme of "mineral composition + TOC content"; Step 2: determine the target layer for which marine shale lithofacies division is required, and obtain well logging data of the target layer; Determine multiple depths at equal intervals on the target layer and obtain the mineral content and organic matter content of the core at each depth; Step 3: For each logging curve in the target layer logging data, a correlation analysis is performed with the mineral content and the organic matter content to obtain the correlation coefficient between each logging curve and the mineral content and the organic matter content, and the logging curve is selected according to the correlation coefficient; Step 4: Select data from target layers of multiple wells and construct a sample data set after normalization. Step 5: Optimize the RBF neural network using the K-means clustering algorithm, take the sample data as input, and the mineral and organic matter contents at the depth corresponding to the sample data as output, and train the prediction models of mineral content and organic matter content respectively; Step 6: Perform detailed lithology division based on the output data of the prediction model.

2. The marine shale lithofacies logging fine identification method according to claim 1 is characterized in that: The marine shale lithofacies division scheme determined in step 1 is: Marine shale lithofacies are divided into two categories based on the TOC content of 2%: organic-rich shale with a TOC content ≥ 2% and organic-poor shale with a TOC content < 2%; According to the main mineral composition of shale, the contents of carbonate minerals, clay minerals and siliceous minerals are used as the three end members of the triangle chart; calcite and dolomite represent carbonate minerals; feldspar and quartz represent siliceous minerals; montmorillonite, kaolinite, illite-montmorillonite mixed layer, chlorite-montmorillonite mixed layer, illite and chlorite represent clay minerals; The shale phases with a content exceeding 50% of the dividing line are named calcareous shale phase, siliceous shale phase, and clay shale phase, and the shale phases with a mineral content not exceeding 50% are named mixed shale phase.

3. The marine shale lithofacies logging fine identification method according to claim 1 is characterized in that: In step 2, the logging data of the target layer includes various logging curves including density, natural gamma, porosity, shear wave time difference, potassium element, uranium element, uranium-free gamma, and water saturation.

4. The marine shale lithofacies logging fine identification method according to claim 1 is characterized in that: In step 2, the mineral content of the core is obtained by whole-rock analysis of the core using X-ray diffraction, and the total organic carbon content of the core is obtained by geochemical analysis.

5. The marine shale lithofacies logging fine identification method according to claim 1 is characterized in that: In step three, SPSS statistical analysis software was used for correlation analysis.

6. The marine shale lithofacies logging fine identification method according to claim 1 is characterized in that: In step three, by presetting a threshold, the logging curves with correlation coefficients greater than the threshold are retained for subsequent steps.

7. The marine shale lithofacies logging fine identification method according to claim 1 is characterized in that: In step 4, the sample data set constructed is l j = {l 1j , l 2j , … , l ij , … , l Nj }; where l ij It represents the value of the i-th logging curve at the j-th depth, i=1~N, j=1~M, N is the number of logging curves, and M is the number of multiple depths determined at equal intervals on the target layer.

8. A marine shale lithofacies logging fine identification device, characterized in that: include: Shale lithofacies division module, used to determine the marine shale lithofacies division scheme of "mineral composition + TOC content"; A data acquisition module, used to determine the target layer for which marine shale lithofacies division is required, and to obtain logging data of the target layer; Determine multiple depths at equal intervals on the target layer and obtain the mineral content and organic matter content of the core at each depth; The logging curve screening module is used to perform correlation analysis on each logging curve in the target layer logging data with the mineral content and organic matter content, obtain the correlation coefficient between each logging curve and the mineral content and organic matter content, and select the logging curve according to the correlation coefficient; The sample data construction module is used to select data from target layers of multiple wells and construct a sample data set after normalization processing; The model training module is used to optimize the RBF neural network using the K-means clustering algorithm, taking the sample data as input and the mineral and organic matter contents at the corresponding depth of the sample data as output, and respectively training to obtain the prediction models of mineral content and organic matter content; The lithofacies division module is used to perform detailed lithofacies division based on the output data of the prediction model.

9. A marine shale lithofacies logging fine identification device, characterized in that: It comprises a processor and a memory for storing processor executable instructions, wherein the instructions, when executed by the processor, implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images