A method and device for classifying pore structure using fractal dimension of well logging

Abstract

The present invention provides a method, device, storage medium, and computer equipment for classifying pore structure using fractal logging dimensions. The method comprises obtaining a capillary pressure curve of a reservoir core and classifying the reservoir pore structure based on the capillary pressure curve; extracting logging data for the well section corresponding to the capillary pressure curve of the core, and determining sensitive parameter logging curves corresponding to different pore structure types based on the logging data; calculating the correlation fractal dimension of the sensitive parameter logging curves corresponding to different pore structure types using a correlation fractal dimension calculation method, wherein the correlation fractal dimension is used to quantitatively describe the complexity of the pore structure; using the correlation fractal dimension of the sensitive parameter logging curves corresponding to different pore structure types as sample data, establishing a pore structure classification and recognition training model based on a cluster analysis algorithm; and performing machine learning using the pore structure classification and recognition training model to classify the pore structure of the entire well section.

Description

Technical Field

[0001] The present invention relates to the field of petroleum and geological exploration technology, and in particular to a method and device for classifying pore structure types by using fractal dimension of well logging data and machine learning, as well as corresponding storage media and computer equipment. Background Art

[0002] Pore ​​structure refers to the geometry, size, distribution, interconnectivity, and configuration of pores and throats within a rock. Pore structure primarily reflects the combination of various pore types and the connecting throats within a reservoir, providing an overall picture of pore and throat development. The microscopic pore structure characteristics of reservoir rocks are the primary factors influencing the storage capacity of reservoir fluids (oil, gas, and water) and the extraction of oil and gas resources. Therefore, understanding the pore structure characteristics of reservoir rocks is crucial for increasing oil and gas reservoir production and improving oil and gas recovery rates.

[0003] Hu Yong et al. (Hu Yong, Zhu Huayin, Wan Yujin, et al. Pore structure and gas-water seepage characteristics of Daqing volcanic rocks [J]. Journal of Southwest Petroleum University (Natural Science Edition), 2007, 29(5): 63-65, 89. 2007) used CT, mercury intrusion, nuclear magnetic resonance and thin sections to classify the pore structure of volcanic rocks, and divided the reservoirs into three types: porous type, fracture type and dense type, and studied the gas-water seepage characteristics of different types. Guan Liqun et al. (Guan Liqun, Qu Hongjun, Hu Chunhua, et al. Study on the relationship between reservoir heterogeneity and oil content of Chang 6 oil layer group in H area of ​​Ansai Oilfield. Lithologic Reservoirs, 2010, 22(3): 26-30, 37.) studied the pore structure types of Ansai Oilfield based on cast thin sections and mercury intrusion data, and divided the pores into primary intergranular pores, dissolved intragranular pores, micropores and microcracks. Wang Ruifei et al. (Wang Ruifei, Chen Mingqiang, Sun Wei. Classification and evaluation of microscopic pore structure of ultra-low permeability sandstone reservoirs [J]. Acta Geoscientia Sinica, 2008, 29(2): 213-220.) used conventional mercury injection measurement methods to evaluate the classification of microscopic pore structure of ultra-low permeability sandstone reservoirs. The comprehensive evaluation parameters of reservoir pore structure and mercury desaturation parameters were established, and the mercury injection test samples were processed separately. The classification results were relatively consistent with the actual development characteristics of the oil field. Li Chaoliu et al. (Li Chao-Liu, Zhou Can-Can, Li Xia, et al. A novel model for assessing the pore structure of tightsands and its application [J]. Applied Geophysics, 2010, 7(3): 283-291.) constructed a comprehensive pore structure evaluation index composed of porosity, maximum connected pore radius and sorting coefficient; Su Junlei (Su Junlei, Sun Jianmeng, Wang Tao, et al. Improved method for evaluating reservoir pore structure using nuclear magnetic resonance logging data [J]. Journal of Jilin University (Earth Science Edition), 2011, 41(Supplement 1), 380-386.) combined nuclear magnetic resonance and oil testing data to screen out reservoir classification evaluation parameters such as effective porosity, absolute permeability, displacement pressure, pore throat mean, and sorting coefficient; Chen Jing et al. (Chen Jing, Wang Guiwen, Zhou Zhenglong, et al. Classification and evaluation of pore structure in tight oil reservoirs and analysis of its genesis [J]. Progress in Geophysics, 2017, 32(03): 1095-1105.) established a comprehensive diagenetic coefficient and T2 geometric mean considering compaction, cementation diagenesis and micropore content to quantitatively characterize different types of pore structures; Ni Guohui et al. (Ni Guohui, Guo Haifeng, Xu Xing, et al. Well logging identification and classification evaluation of complex pore structure in carbonate rocks - a case study of H formation in an oil field in the Middle East [J]. Journal of Oil and Gas, 2014, 36(01): 60-65+6.) used electrical imaging and nuclear magnetic logging data to establish a pore structure and reservoir type identification chart for the H area; Li Xiaofeng and Peng Shimi (Li Xiaofeng, Peng Shimi. Microscopic pore structure characteristics and classification evaluation of the reservoir of Kangcun Formation in Dawanqi Oilfield [J]. Journal of Petroleum and Natural Gas, 2011, 33(06): 26-31+6.) used cast thin sections, scanning electron microscopy, mercury injection testing and other analytical data combined with principal factor analysis and cluster analysis to divide the pore structure of the reservoir of Kangcun Formation into four types basic types; Sun Junchang et al. (Sun Junchang, Zhou Hongtao, Guo Hekun, et al. Fractal geometry description of microscopic heterogeneity of complex reservoir rocks [J]. Journal of Wuhan Polytechnic Institute, 2009, 28 (3): 42-46.) calculated the fractal dimension using the capillary pressure-mercury saturation model and used it to evaluate reservoir heterogeneity; Zhao Jie et al. (Zhao Jie, Jiang Yizhong, Wang Weinan, et al. Experimental study on determining rock pore structure using nuclear magnetic resonance technology [J]. Well Logging Technology, 20 03, 27(3): 185-188.) The relationship between the conversion model coefficient and the porosity and permeability ratio was established by comparing the nuclear magnetic resonance T2 spectrum with the mercury intrusion pore size distribution, and the influence of paramagnetic materials on the conversion coefficient was considered and a method for determining the conversion model coefficient was proposed; Tan Maojin et al. (Tan Maojin, Zhao Wenjie. Evaluation of complex lithologic reservoirs such as carbonate rocks using nuclear magnetic resonance logging data [J]. 2006, 21(2): 489-493.) used the nuclear magnetic resonance logging T2 distribution to determine the reservoir space type and reservoir effectiveness; Liu Wei et al. (Liu Wei, Xiao Zhongxiang, Yang Siyu, et al. Comparative study of methods for evaluating reservoir pore structure using nuclear magnetic resonance (NMR) logging data [J]. Petroleum Geophysical Exploration, 2009, 44(6): 773-778.) compared the application effects of the three pore component ratio, similarity comparison method, average saturation error minimum method and Swanson parameter method in nuclear magnetic resonance pore structure evaluation.

[0004] In summary, the evaluation of reservoir pore structure mainly focuses on four types of methods: 1. Statistical analysis of thin section, porosity, mercury injection and other data. This method is based on mathematical statistics. The disadvantage is that it requires a lot of experimental data and cannot be continuously quantitatively evaluated or predicted; 2. Based on the core capillary pressure experiment, the correlation between parameters such as displacement pressure and roar channel mean and the logging curve is established, and then the logging curve is used to continuously evaluate the pore structure; the difficulty of this method is that it is generally difficult to establish a fitting relationship with a high compliance rate; 3. Using scanning electron microscopy and other methods combined with the pore network model to intuitively analyze the pore structure. This method intuitively reflects the pore structure, but it is difficult to continuously process the entire well section and predict unknown well sections; 4. Using nuclear magnetic resonance logging data for pore structure evaluation. This method is the most direct logging method to reflect the pore structure, but nuclear magnetic resonance logging is expensive and its wide application is subject to certain restrictions. Summary of the Invention

[0005] The present invention generates a training model based on the logging fractal dimension of the logging curves of sensitive parameters of different pore structure types, and uses the least-nearest neighbor algorithm (KNN) to perform machine learning on the training model to carry out pore structure classification and evaluation, achieving the purpose of quantitative and continuous pore structure classification and evaluation using conventional data.

[0006] According to one embodiment of the present invention, a method for classifying pore structures using fractal dimension of well logging is provided, which mainly includes the following steps:

[0007] S100, obtaining a capillary pressure curve of a reservoir core, and classifying the reservoir pore structure according to the capillary pressure curve of the core;

[0008] S300, extracting logging data of the well section corresponding to the capillary pressure curve of the core, and determining sensitive parameter logging curves corresponding to different pore structure types based on the logging data;

[0009] S500, using a correlation fractal dimension calculation method to calculate the correlation fractal dimension of sensitive parameter logging curves corresponding to different pore structure types, wherein the correlation fractal dimension is used to quantitatively describe the complexity of the pore structure;

[0010] S700, using the correlation dimension and fractal dimension of the sensitive parameter logging curves corresponding to different pore structure types as sample data, a pore structure classification and recognition training model is established based on a cluster analysis algorithm;

[0011] S900 uses a pore structure classification and recognition training model for machine learning to classify the pore structure of the entire well section.

[0012] According to an embodiment of the present invention, in step 100, the pore structure is classified according to the capillary pressure curve shape, displacement pressure, mercury saturation median pressure, median pore radius, and maximum unsaturated pore volume.

[0013] According to one embodiment of the present invention, step 300 includes the following steps:

[0014] Extract the logging data of the well section corresponding to the core capillary pressure curve and make the logging data distribution histogram of different pore structure types;

[0015] According to the distribution histogram of logging data of different pore structure types, the sensitive parameter logging curves corresponding to different pore structure types are determined.

[0016] According to one embodiment of the present invention, the sensitive parameter logging curve includes at least one of a gamma curve, a spontaneous potential curve, an acoustic wave curve, a neutron curve, a density curve, a deep resistivity curve, and a shallow resistivity curve.

[0017] According to one embodiment of the present invention, step 500 includes the following steps:

[0018] S501, loading the logging data file, selecting the sensitive parameter logging curve for correlation dimension fractal dimension calculation;

[0019] S502, inputting the starting depth and ending depth of the well section for calculating the correlation dimension fractal dimension;

[0020] S503, determining the logging data corresponding to the starting depth and the ending depth according to the depth sampling rate of the logging curve, and determining the number M of logging data in the calculation well section;

[0021] S504, input vector dimension N and the number of data shifts T when constructing the embedding space;

[0022] S505, assigning values ​​to each vector to form a vector space, thereby constructing an embedding space;

[0023] S506, calculating the Euclidean distance between any two vectors, and finding the maximum distance and the minimum distance from the calculation results;

[0024] S507, determining different scales according to the given number of statistical scales and the maximum distance and the minimum distance;

[0025] S508, calculating the ratio of the number of distances meeting the conditions to the total number of distances at different scales;

[0026] S509, calculating the correlation dimension fractal dimension D using all scales and their statistical ratios;

[0027] S510, check the number of well sections for calculating the correlation dimension fractal dimension. If the input well section has not been calculated, return to step S503 and repeat the above calculation process; if the input well section has been calculated, output the calculation result.

[0028] According to one embodiment of the present invention, in the above step 700, the input data of the pore structure classification and identification training model is the correlation dimension of the sensitive parameter logging curve, the supervision data is the pore structure type, and the output data is the number of data clusters of different pore structure types.

[0029] According to one embodiment of the present invention, in the above step 900, based on the minimum neighbor algorithm, machine learning is performed using the pore structure classification and recognition training model to classify the pore structure of the entire well section.

[0030] In addition, the present invention also provides a well logging fractal dimension pore structure classification device, which includes:

[0031] The sample structure classification module is used to obtain the capillary pressure curve of the reservoir core and classify the reservoir pore structure according to the capillary pressure curve of the core;

[0032] The sensitive parameter logging curve determination module is used to extract the logging data of the well section corresponding to the core capillary pressure curve, and determine the sensitive parameter logging curves corresponding to different pore structure types based on the logging data;

[0033] A correlation dimension calculation module is used to calculate the correlation dimension fractal dimension of sensitive parameter logging curves corresponding to different pore structure types using a correlation fractal dimension calculation method, wherein the correlation dimension fractal dimension is used to quantitatively describe the complexity of the pore structure;

[0034] The training model construction module is used to use the correlation dimension and fractal dimension of the sensitive parameter logging curves corresponding to different pore structure types as sample data and establish a pore structure classification and recognition training model based on the cluster analysis algorithm;

[0035] The pore structure classification module is used to classify the pore structure of the entire well section using the pore structure classification and recognition training model.

[0036] In addition, the present invention also provides a storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the well logging fractal dimension pore structure classification method as described above.

[0037] In addition, the present invention also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the steps of the well logging fractal dimension pore structure classification method described above are implemented.

[0038] Compared with the prior art, one or more embodiments of the above solutions may have the following advantages or beneficial effects:

[0039] The fractal dimension pore structure classification method for well logging curves provided by the present invention utilizes the fractal dimension of sensitive parameter well logging curves to determine the fractal dimension range for each pore structure type. Based on this, cluster analysis is used to generate a training model, and machine learning is used to classify and evaluate reservoir pore structure. The fractal dimension pore structure classification method for well logging curves provided by the present invention fully considers the response of pore structure to macroscopic well logging curves, reducing the computational effort required for fractal dimension calculations of non-sensitive parameter well logging curves. Furthermore, the fractal dimension pore structure classification method utilizes machine learning to achieve more reliable classification results. The provided fractal dimension pore structure classification method for well logging curves has a strong theoretical basis, is simple to operate, and offers high accuracy. It can be used for quantitative evaluation of reservoir pore structure and is highly practical.

[0040] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the structures particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0042] Figure 1 This is a flow chart of a method for classifying pore structures based on fractal dimensions of well logging curves according to an embodiment of the present invention;

[0043] Figure 2 Schematic diagram of capillary pressure curves of four pore structures according to an embodiment of the present invention;

[0044] Figure 3 Schematic diagram of logging curves of well logging sensitivity parameters corresponding to different pore structure types according to an embodiment of the present invention;

[0045] Figure 4 Schematic diagram of the distribution range of correlation dimensions of logging curves of well logging sensitive parameters corresponding to different pore structure types according to an embodiment of the present invention;

[0046] Figure 5 A pore structure classification and recognition training model for four pore structures established in an embodiment of the present invention;

[0047] Figure 6 Schematic diagram of the results obtained by the pore structure classification method using the fractal dimension of well logging curves according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the accompanying drawings to clearly and completely describe the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0049] This invention discloses a method for quantitatively evaluating pore structure types using well logging data. This method generates a training model based on the fractal dimension of sensitive logging curves and employs machine learning to quantitatively evaluate pore structure types. This method improves reservoir logging evaluation technology and expands the geological application scope of well logging data, belonging to the field of petroleum and geology. The method primarily includes the following steps:

[0050] S1. Use core capillary pressure curve to classify reservoir pore structure.

[0051] The pore structure types are divided into four categories according to pore structure parameters such as capillary pressure curve morphology, displacement pressure (Pd), mercury saturation median pressure (Pc50), median pore radius (Rc50), and maximum unsaturated pore volume (Smax).

[0052] Type I: Mercury injection curve (injection curve) is located at Figure 2 At the lower position in A, the curve is smoother and has the characteristics of lower displacement pressure, larger median roar radius and higher mercury saturation.

[0053] Type II: Mercury injection curve (injection curve) is located at Figure 2 At the lower middle part of B, the curve is smoother and has the characteristics of medium displacement pressure, medium median throat radius and high mercury saturation.

[0054] Type III: Mercury injection curve (injection curve) is located at Figure 2 In the upper middle part of C, the curve is inclined without a platform, and the curve shows the characteristics of higher displacement pressure, higher median throat radius and lower mercury saturation.

[0055] Type IV: Mercury injection curve (injection curve) is located at Figure 2 At the upper part of D, the curve is inclined without a platform, and the curve is characterized by high displacement pressure, high median throat radius and low mercury saturation.

[0056] Displacement pressure (Pd) refers to the capillary pressure of the largest connected pore in the pore system. That is, the intersection of the tangent line along the flat part of the capillary pressure curve and the vertical axis is the Pd value.

[0057] Mercury saturation median pressure (Pc50) refers to the capillary pressure of the injection curve corresponding to the mercury saturation of 50%;

[0058] The median pore radius (Rc50) refers to the pore radius corresponding to the median pressure of mercury saturation (Pc50);

[0059] The maximum unsaturated pore volume (Smax) refers to the percentage of pore volume invaded by the watermark when the pressure of the injected mercury reaches the maximum pressure of the instrument.

[0060]

[0061] S2. Extract the logging data of the well section corresponding to the capillary pressure curve.

[0062] The logging data of the deep well section corresponding to the capillary pressure curve are extracted, including acoustic wave (AC), natural gamma (GR), spontaneous potential (SP), wellbore (CAL), photoelectric interface index (PE), neutron (CNL), density (DEN) and deep and shallow resistivity (RD, RS) curve data.

[0063] S3. Extract logging curves of sensitive parameters of different pore structure types.

[0064] Based on the logging curve data of S2, the logging data distribution histograms of different pore structure types were made and the logging curves of sensitive parameters were extracted.

[0065] Different pore structure types are reflected in different shapes and values ​​of logging curves, that is, different sensitive parameter logging curves. Figure 3 In the PE value distribution histogram for different pore structures, the PE value distribution concentration range for Type II pore structure is significantly different from that of the other three types, indicating that Type II pore structure is sensitive to the PE curve, and so on. The purpose of extracting sensitive parameter logging curves is to calculate the fractal dimension of the corresponding sensitive parameter logging curves, providing a basis for automatic full-well pore structure classification and continuous processing.

[0066] S4. Calculate the correlation dimension of the sensitive parameter logging curve.

[0067] Using the calculation principle of correlation fractal dimension, the correlation fractal dimension of sensitive parameter logging curves corresponding to different pore structure types in S3 is calculated as sample data.

[0068] The pore space of reservoir rocks exhibits excellent fractal characteristics, and its fractal dimension can be used to quantitatively describe the complexity of the pore structure. The larger the fractal dimension, the more complex the reservoir rock's pore structure, while the smaller the fractal dimension, the more homogeneous the reservoir rock and the simpler the pore structure.

[0069] Correlation dimension is a fractal dimension commonly used to calculate the fractal dimension of well logging curves. The basic principle is to use the embedding space method. The specific calculation steps are as follows:

[0070] S4.01 loads the logging data file and selects the logging curve for correlation dimension calculation;

[0071] S4.02 inputs the starting depth and ending depth of the well section for calculating the correlation digit;

[0072] S4.03 determines the logging data corresponding to the starting depth and the ending depth according to the depth sampling rate of the logging curve, and determines the number M of logging data in the calculation well section;

[0073] S4.04 Input vector dimension N and the number of data shifts T when constructing the embedding space;

[0074] S4.05 assigns a value to each vector to form a vector space, thereby constructing an embedding space;

[0075] S4.06 Calculate the Euclidean distance R between any two vectors ij , and find the maximum distance R max and the minimum distance R min ;

[0076] S4.07 Given the number of statistical scales, according to the maximum distance R max and the minimum distance R min Determine different scales ε;

[0077] S4.08 Calculate the R at different scales ij The ratio of the number of distances less than ε to the total number of distances;

[0078] S4.09 calculates the correlation dimension D based on the proportion of all scales and their statistics;

[0079] S4.10 checks the number of well sections for calculating the correlation dimension. If the input well section has not been calculated, returns to step 3 and repeats the above calculation process; if the input well section has been calculated, outputs the calculation result.

[0080] S5. Generate a pore structure classification and recognition training model.

[0081] According to S1 and S3, make S4 different fractal dimension intersection diagrams to determine the limit range of the fractal dimension of each type of pore structure. Figure 4 Since the cross-plot classification boundaries cannot completely separate the pore structures, a cluster analysis algorithm is required to establish a pore structure classification and recognition training model. The input data of the training model is the correlation dimension of the sensitive parameter logging curves of the continuous well section, the supervision data is the pore structure type code of the corresponding depth position in S1 (in this embodiment, PORETYPE=1, 2, 3, 4 represent pore structure types I, II, III, and IV respectively), and the output data is the number of data clusters of different pore structure types, that is, Figure 5 Different data clusters represent different pore structure types.

[0082]

[0083] In the above table, D RD / AC 、D PE 、D DEN 、D CNL 、D RD / RS They represent the correlation dimension of RD / AC curve, the correlation dimension of PE curve, the correlation dimension of DEN curve, the correlation dimension of CNL curve and the correlation dimension of RD / RS curve respectively.

[0084] S6. Conduct machine learning to carry out classification and evaluation of pore structure in the entire well section.

[0085] Based on the pore structure classification and identification training model determined by S5, the KNN algorithm is used for machine learning to carry out the pore structure classification evaluation of the entire well section of the unknown well.

[0086] The training model in S5 uses data with known pore structure types as supervision. To classify and evaluate the pore structure of the entire blind well section, a least-neighbor (KNN) algorithm is used for machine learning, outputting pore structure classification codes for continuous well sections. KNN is a theoretically mature method. The idea is that if a sample's k most similar (i.e., closest) samples in feature space belong to a certain category, then the sample also belongs to that category. The distance between this point and all other points is calculated, and the k points closest to the point are selected. The point with the largest proportion of these k points is then considered to belong to that category.

[0087] Through the KNN algorithm, the connection between the training model and the pore structure classification curve (PORETYPE) can be established. Figure 6 As shown in the figure, the last PORTYPE is the classification result curve.

[0088] Example 2

[0089] In addition, this embodiment also provides a well logging fractal dimension pore structure classification device, which mainly includes:

[0090] The sample structure classification module is used to obtain the capillary pressure curve of the reservoir core and classify the reservoir pore structure according to the capillary pressure curve of the core;

[0091] The sensitive parameter logging curve determination module is used to extract the logging data of the well section corresponding to the core capillary pressure curve, and determine the sensitive parameter logging curves corresponding to different pore structure types based on the logging data;

[0092] A correlation dimension calculation module is used to calculate the correlation dimension fractal dimension of sensitive parameter logging curves corresponding to different pore structure types using a correlation fractal dimension calculation method, wherein the correlation dimension fractal dimension is used to quantitatively describe the complexity of the pore structure;

[0093] The training model construction module is used to use the correlation dimension and fractal dimension of the sensitive parameter logging curves corresponding to different pore structure types as sample data and establish a pore structure classification and recognition training model based on the cluster analysis algorithm;

[0094] The pore structure classification module is used to classify the pore structure of the entire well section using the pore structure classification and recognition training model.

[0095] Example 3

[0096] In addition, this embodiment provides a storage medium storing a computer program. When the computer program is executed by a processor, the steps of the well logging fractal dimension pore structure classification method described in the first embodiment are implemented.

[0097] Example 4

[0098] In addition, this embodiment provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the steps of the well logging fractal dimension pore structure classification method as described in Example 1 are implemented.

[0099] The present invention discloses a method for quantitatively evaluating pore structure types using well logging data. A training model is generated based on the fractal dimension of sensitive parameter logging curves, and machine learning is used to quantitatively evaluate pore structure types. This approach improves reservoir logging evaluation technology and expands the geological application scope of logging data, belonging to the field of petroleum and geology. By extracting logging sensitive parameters of different pore structure types, the present invention establishes a training model for machine learning to identify pore structure types. The extracted sensitive parameters avoid the influence of fluid and other information on logging curves, reducing the computational effort required for non-sensitive parameter logging correlation dimensions, thus facilitating the precise quantitative evaluation of pore structure in heterogeneous formations during oil and gas exploration.

[0100] It should be noted that the method of the embodiment of the present invention can be performed by a single device, such as a computer or server. The method of this embodiment can also be applied in a distributed scenario, where multiple devices cooperate to perform the method. In such a distributed scenario, one of the multiple devices may only perform one or more steps of the method of the embodiment of the present invention, and the multiple devices will interact with each other to complete the method.

[0101] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0102] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0103] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0104] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing module, or each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.

[0105] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0106] Although the embodiments disclosed herein are as described above, the contents described herein are merely embodiments for facilitating understanding of the present invention and are not intended to limit the present invention. Any person skilled in the art may make any modifications and variations in the form and details of the embodiments without departing from the spirit and scope of the present invention. However, the scope of protection of the present invention shall remain subject to the scope defined by the appended claims.

Claims

1. A method for classifying pore structures using fractal dimension of well logging, characterized in that: include: S100, obtaining a capillary pressure curve of a reservoir core, and classifying the reservoir pore structure according to the capillary pressure curve of the core; S300, extracting logging data of the well section corresponding to the capillary pressure curve of the core, and determining sensitive parameter logging curves corresponding to different pore structure types based on the logging data; S500, using a correlation fractal dimension calculation method to calculate the correlation fractal dimension of sensitive parameter logging curves corresponding to different pore structure types, wherein the correlation fractal dimension is used to quantitatively describe the complexity of the pore structure; S700, using the correlation dimension and fractal dimension of the sensitive parameter logging curves corresponding to different pore structure types as sample data, a pore structure classification and recognition training model is established based on a cluster analysis algorithm; S900, which uses a pore structure classification and recognition training model for machine learning to classify the pore structure of the entire well section; In step 700, the input data of the pore structure classification and identification training model is the correlation dimension of the sensitive parameter logging curve, the supervision data is the pore structure type, and the output data is the number of data clusters of different pore structure types.

2. The well logging fractal dimension pore structure classification method according to claim 1, characterized in that: In step 100, the pore structure is classified according to the capillary pressure curve shape, displacement pressure, mercury saturation median pressure, median pore radius, and maximum unsaturated pore volume.

3. The well logging fractal dimension pore structure classification method according to claim 1, characterized in that: The step 300 includes the following steps: Extract the logging data of the well section corresponding to the core capillary pressure curve and make the logging data distribution histogram of different pore structure types; According to the distribution histogram of logging data of different pore structure types, the sensitive parameter logging curves corresponding to different pore structure types are determined.

4. The method for classifying pore structures based on well logging fractal dimension according to claim 3, characterized in that: The sensitive parameter logging curve includes at least one of a gamma curve, a natural potential curve, an acoustic wave curve, a neutron curve, a density curve, a deep resistivity curve, and a shallow resistivity curve.

5. The well logging fractal dimension pore structure classification method according to claim 1, characterized in that: The step 500 includes the following steps: S501, loading the logging data file, selecting the sensitive parameter logging curve for correlation dimension fractal dimension calculation; S502, inputting the starting depth and ending depth of the well section for calculating the correlation dimension fractal dimension; S503, determining the logging data corresponding to the starting depth and the ending depth according to the depth sampling rate of the logging curve, and determining the number M of logging data in the calculation well section; S504, input vector dimension N and the number of data shifts T when constructing the embedding space; S505, assigning values ​​to each vector to form a vector space, thereby constructing an embedding space; S506, calculating the Euclidean distance between any two vectors, and finding the maximum distance and the minimum distance from the calculation results; S507, determining different scales according to the given number of statistical scales and the maximum distance and the minimum distance; S508, calculating the ratio of the number of distances meeting the conditions to the total number of distances at different scales; S509, calculating the correlation dimension fractal dimension D based on all scales and their statistical ratios; S510, check the number of well sections for calculating the correlation dimension fractal dimension. If the input well section has not been calculated, return to step S503 and repeat the above calculation process; if the input well section has been calculated, output the calculation result.

6. The method for classifying pore structures based on well logging fractal dimension according to claim 1, characterized in that: In step 900, based on the minimum neighbor algorithm, machine learning is performed using the pore structure classification and recognition training model to classify the pore structure of the entire well section.

7. A well logging fractal dimension pore structure classification device, characterized in that: include: The sample structure classification module is used to obtain the capillary pressure curve of the reservoir core and classify the reservoir pore structure according to the capillary pressure curve of the core; The sensitive parameter logging curve determination module is used to extract the logging data of the well section corresponding to the core capillary pressure curve, and determine the sensitive parameter logging curves corresponding to different pore structure types based on the logging data; A correlation dimension calculation module is used to calculate the correlation dimension fractal dimension of sensitive parameter logging curves corresponding to different pore structure types using a correlation fractal dimension calculation method, wherein the correlation dimension fractal dimension is used to quantitatively describe the complexity of the pore structure; The training model construction module is used to use the correlation dimension and fractal dimension of the sensitive parameter logging curves corresponding to different pore structure types as sample data and establish a pore structure classification and recognition training model based on the cluster analysis algorithm; The pore structure classification module is used to classify the pore structure of the entire well section using the pore structure classification and recognition training model.

8. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the well logging fractal dimension pore structure classification method according to any one of claims 1 to 6 are implemented.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the steps of the well logging fractal dimension pore structure classification method according to any one of claims 1 to 6 are implemented.

Images