A diagenetic facies identification method based on geological constraints and well logging parameter fusion clustering

Abstract

The present invention relates to a diagenetic phase identification method based on geological constraint-based logging parameter fusion clustering, including: collecting logging data of the entire well section of the target well and normalizing it; fusing and calculating the processed logging data based on geological knowledge constraints to obtain fused logging data; clustering the fused logging data using an unsupervised machine learning algorithm and then combining them to obtain initial logging phase types; screening the initial logging phase types based on geological law constraints and then performing intersection and merging processing to obtain target logging phase types; selecting a number of sample points in the core section of the target well and marking them for diagenetic phases, constructing a diagenetic phase data set and calibrating and verifying the target logging phase types, obtaining the diagenetic phase types corresponding to each target logging, and completing the diagenetic phase identification of the entire well section. The present invention comprehensively applies various logging parameters under geological constraints to achieve accurate identification and prediction of diagenetic phases under low sample quantity conditions.

Description

Technical Field

[0001] The present invention relates to the field of oil and gas reservoirs and machine learning technology, and in particular to a diagenetic phase identification method based on geological constraint-based logging parameter fusion clustering. Background Art

[0002] Diagenesis controls the evolution of the physical properties of oil and gas reservoirs, and diagenetic phases are a comprehensive reflection of the evolutionary results of diagenesis. Therefore, accurate identification and prediction of diagenetic phases are of great significance for oil and gas reservoir evaluation and efficient exploration and development of oil and gas resources. Traditional diagenetic phase research methods rely on the analysis and testing of samples from limited cored well sections, and it is difficult to accurately predict the diagenetic phases of the entire well section. In recent years, big data and machine learning technologies have become an important means of oil and gas reservoir research. The relationship between diagenetic phases and well logging data can be used to identify and predict diagenetic phases of the entire well section using machine learning algorithms. At present, supervised learning methods are mostly used for diagenetic phase logging identification, including: decision trees, random forests, XGboost, BP neural networks, etc. Existing technologies lack integration with geological knowledge, may ignore or overuse some logging information, and in practical applications, it is often difficult to meet the large sample size requirements of supervised machine learning algorithms, affecting the accuracy of diagenetic phase identification results. Therefore, it is urgent to propose an unsupervised machine learning method that comprehensively considers logging data based on geological constraints to achieve accurate identification and prediction of diagenetic facies in study areas with fewer sample points. Summary of the invention

[0003] The present invention aims to solve the deficiencies of the prior art and proposes a diagenetic phase identification method based on geological constraint-based logging parameter fusion clustering. The logging parameters are fused based on geological constraints, and the initial logging phase is obtained by clustering the fused logging parameters. The target logging phase is obtained based on geological constraints and intersection and merging processing. Based on the diagenetic phase marking results of the coring section sample points, the corresponding relationship between the target logging phase and the diagenetic phase is determined to achieve the identification and prediction of the diagenetic phase.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] A diagenetic facies identification method based on geological constraints and logging parameter fusion clustering specifically comprises the following steps:

[0006] Step 1: collecting logging data of the entire well section of the target well and normalizing the logging data;

[0007] Step 2: Based on geological knowledge constraints, the processed logging data is fused and calculated to obtain fused logging data;

[0008] Step 3: clustering the fused logging data respectively using an unsupervised machine learning algorithm, and then combining the clustered fused logging data to obtain the initial logging phase type of the entire well section;

[0009] Step 4: Screening the initial logging phase type based on geological law constraints, and performing intersection and merging processing on the screened results to obtain the target logging phase type of the entire well section, wherein the geological law constraints include diagenesis and pore response characteristics;

[0010] Step 5, select a number of sample points in the core section of the target well, mark the diagenetic facies of the sample points through cast thin section analysis, and construct a diagenetic facies data set;

[0011] Step 6: Calibrate and verify the target logging facies type based on the diagenetic facies data set, obtain the diagenetic facies type corresponding to each target logging, and complete the diagenetic facies identification of the entire well section.

[0012] Optionally, the types of logging data in step 1 include: natural gamma ray GR, spontaneous potential SP, shallow resistivity RS, deep resistivity RD, acoustic time difference AC, compensated neutron CNL, and density DEN.

[0013] Optionally, the method for normalizing the well logging data in step 1 is:

[0014]

[0015] Where, X nor represents the normalized logging data, X represents the logging data, and X min Indicates the minimum value of the logging data, X max Indicates the maximum value of the logging data.

[0016] Optionally, in step 2, performing fusion calculation on the processed logging data based on geological knowledge constraints to obtain fused logging data includes:

[0017] Performing fusion calculation on the normalized logging data based on a preset fusion principle to obtain the fused logging data, wherein the fused logging data includes first fused logging data, second fused logging data, and third fused logging data;

[0018] The fusion principle is set based on geological constraints. The fusion principle of the first fusion logging data is to fuse GR and SP, which are sensitive to mud content and lithology, into parameter GS. The fusion principle of the second fusion logging data is to fuse RS and RD, which are sensitive to cement and fracture responses, into parameter SD. The fusion principle of the third fusion logging data is to fuse AC, CNL and DEN, which are sensitive to reservoir porosity responses, into parameter ACD.

[0019] Optionally, the method for performing the fusion calculation is:

[0020] X G-S =X GRnor -X SPnor

[0021] X S-D =X RSnor +X RDnor

[0022] X A-C-D =X ACnor +X CNLnor -X DENnor

[0023] In the formula, X G-S represents the fusion value of the fusion parameter GS, X S-D represents the fusion value of the fusion parameter SD, X A-C-D represents the fusion value of the fusion parameter ACD, X GRnor represents the normalized value of GR, X SPnor represents the normalized value of SP, X RSnor represents the normalized value of RS, X RDnor represents the normalized value of RD, X ACnor represents the normalized value of AC, X CNLnor represents the normalized value of CNL, X DENnor Indicates the normalized value of DEN.

[0024] Optionally, in step 3, clustering the fused logging data respectively using an unsupervised machine learning algorithm, and then combining the clustered fused logging data to obtain the initial logging phase type of the entire well section includes:

[0025] K-means clustering algorithm was used to cluster each type of fused logging data into three categories;

[0026] Based on the clustering combination results of three types of fused logging data, 27 types of initial logging facies were obtained.

[0027] Optionally, in step 4, the initial logging phase type is screened based on geological law constraints, and the screened results are subjected to intersection and merging processing to obtain the target logging phase type of the entire well section, including:

[0028] The initial logging phase types are deleted based on a preset screening mechanism to obtain 16 types of initial logging phases after screening, wherein the screening mechanism is set based on geological law constraints, and the screening mechanism includes: deleting initial logging phases containing mutually opposing diagenesis, and deleting initial logging phases containing mutually contradictory diagenesis and pore response characteristics;

[0029] Based on the fused logging data of the screened 16 types of initial logging phases, a GS, SD, and ACD ternary cross-plot is constructed, and the ternary cross-plot is merged based on the distribution characteristics to determine the target logging phase type of the entire well section.

[0030] Optionally, the diagenetic phase marking of the sample points by casting thin section analysis in step 5 includes:

[0031] Determine the diagenetic parameters of the sample point by analyzing the casting thin section of the sample point, wherein the diagenetic parameters include apparent compaction rate, cement content, surface ratio, and crack ratio;

[0032] A quantitative division standard of diagenetic phases is determined based on the diagenetic parameters, and the diagenetic phases of the sample points are marked.

[0033] Optionally, the calibration and verification of the target logging facies type based on the diagenetic facies dataset in step 6 includes:

[0034] 70% of the data in the diagenetic facies data set are selected to calibrate the target logging facies type, and the remaining 30% of the data in the diagenetic facies data set are selected to verify the accuracy of the calibration result.

[0035] The present invention has the following beneficial effects:

[0036] The present invention is based on an unsupervised machine learning algorithm, fuses logging parameters based on geological constraints, obtains logging phase types by clustering and merging the fused logging parameters, and realizes the identification and prediction of diagenetic phases based on the correspondence between logging phases and diagenetic phases. Compared with the existing technology, the method of the present invention comprehensively applies logging data under the constraints of geological factors, has low requirements on the number and types of sample points, and has a high accuracy rate for diagenetic phase identification. The method of the present invention is widely applicable to the field of sandstone reservoir research, can effectively support the efficient exploration and development of oil and gas resources, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 This is a flow chart of a diagenetic phase identification method based on geological constraints and well logging parameter fusion clustering according to an embodiment of the present invention;

[0039] Figure 2This is an initial well logging phase classification result diagram of an embodiment of the present invention;

[0040] Figure 3 This is a diagram of initial well logging phase screening results according to an embodiment of the present invention;

[0041] Figure 4 This is a target well logging phase classification result diagram of an embodiment of the present invention;

[0042] Figure 5 This is an accuracy analysis diagram of the diagenetic phase calibration results of Well S1 in an embodiment of the present invention;

[0043] Figure 6 This is a diagram of the diagenetic phase logging identification results of the S1 well in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in 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 creative work are within the scope of protection of the present invention.

[0045] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0046] This embodiment provides a diagenetic facies identification method based on geologically constrained logging parameter fusion clustering, including:

[0047] Step 1: Collect logging data of the entire well section of the target well and perform normalization on the logging data;

[0048] Specifically, the types of logging data include: natural gamma GR, natural potential SP, shallow resistivity RS, deep resistivity RD, acoustic time difference AC, compensated neutron CNL, and density DEN.

[0049] The method for normalizing logging data is:

[0050]

[0051] Where, X nor represents the normalized logging data, X represents the logging data, and X min Indicates the minimum value of the logging data, X max Indicates the maximum value of the logging data.

[0052] Step 2: Based on geological knowledge constraints, the processed logging data is fused and calculated to obtain fused logging data;

[0053] Specifically, based on a preset fusion principle, the normalized logging data is fused and calculated to obtain fused logging data, wherein the fused logging data includes first fused logging data, second fused logging data, and third fused logging data;

[0054] The fusion principle is set based on geological constraints. The first fusion principle of logging data is to fuse GR and SP, which are sensitive to mud content and lithology, into parameter GS. The second fusion principle of logging data is to fuse RS and RD, which are sensitive to cement and fracture response, into parameter SD. The third fusion principle of logging data is to fuse AC, CNL and DEN, which are sensitive to reservoir porosity response, into parameter ACD.

[0055] The method for fusion calculation is:

[0056] X G-S =X GRnor -X SPnor

[0057] X S-D =X RSnor +X RDnor

[0058] X A-C-D =X ACnor +X CNLnor -X DENnor

[0059] Where, X G-S represents the fusion value of the fusion parameter GS, X S-D represents the fusion value of the fusion parameter SD, X A-C-D represents the fusion value of the fusion parameter ACD, X GRnor represents the normalized value of GR, X SPnor represents the normalized value of SP, X RSnor represents the normalized value of RS, X RDnor represents the normalized value of RD, X ACnor represents the normalized value of AC, X CNLnor represents the normalized value of CNL, X DENnor Indicates the normalized value of DEN.

[0060] Step 3: cluster the fused logging data using an unsupervised machine learning algorithm, and then combine the clustered fused logging data to obtain the initial logging phase type of the entire well section;

[0061] Specifically, the K-means clustering algorithm was used to cluster each type of fused logging data into three categories;

[0062] Based on the clustering combination results of three types of fused logging data, 27 types of initial logging facies were obtained.

[0063] Step 4: Screen the initial logging phase types based on geological constraints, and perform intersection and merging processing on the screened results to obtain the target logging phase types of the entire well section;

[0064] Specifically, the initial logging phase types are deleted based on the preset screening mechanism to obtain 16 types of initial logging phases after screening, wherein the screening mechanism is set based on geological law constraints, and the screening mechanism includes: deleting the initial logging phases containing mutually opposing diagenesis, and deleting the initial logging phases containing mutually contradictory diagenesis and pore response characteristics;

[0065] The GS, SD, and ACD ternary cross-plots were constructed based on the fused logging data of the 16 types of initial logging phases after screening. The ternary cross-plots were merged based on the distribution characteristics to determine the target logging phase type of the entire well section.

[0066] Step 5: Select several sample points in the core section of the target well, mark the diagenetic facies of the sample points through cast thin section analysis, and construct a diagenetic facies data set;

[0067] Specifically, the diagenetic parameters of the sample points are determined by analyzing the casting thin sections of the sample points, and the diagenetic parameters include apparent compaction rate, cement content, surface rate, and fracture rate;

[0068] The quantitative division standard of diagenetic facies is determined based on diagenetic parameters, the diagenetic facies of the sample points are marked, and a diagenetic facies dataset is constructed.

[0069] Step 6: Calibrate and verify the target logging facies type based on the diagenetic facies data set, obtain the diagenetic facies type corresponding to each target logging, and complete the diagenetic facies identification of the entire well section.

[0070] Specifically, 70% of the data in the diagenetic facies data set are selected to calibrate the target logging facies type, and the remaining 30% of the data in the diagenetic facies data set are selected to verify the accuracy of the calibration result.

[0071] The following is an example of Well S1 in an oil field. Figure 1-Figure 6 A diagenetic facies identification method based on geological constraints and well logging parameter fusion clustering proposed in this embodiment is specifically described and verified, and specifically includes the following steps:

[0072] Step 1: Collect the logging data of the entire well section of Well S1 and perform normalization on the logging data;

[0073] The types of logging data collected include: natural gamma ray GR, natural potential SP, shallow resistivity RS, deep resistivity RD, acoustic time difference AC, compensated neutron CNL, density DEN;

[0074] The method for normalizing logging data is:

[0075]

[0076] In the formula, X nor represents the normalized logging data, X represents the logging data, and X min Indicates the minimum value of the logging data, X max Indicates the maximum value of the logging data.

[0077] Step 2: Based on geological knowledge constraints, the processed logging data is fused and calculated to obtain fused logging data;

[0078] Specifically, based on a preset fusion principle, the normalized logging data is fused and calculated to obtain fused logging data, wherein the fused logging data includes first fused logging data, second fused logging data, and third fused logging data;

[0079] The fusion principle is set based on geological constraints. The first fusion principle of logging data is to fuse GR and SP, which are sensitive to mud content and lithology, into parameter GS. The second fusion principle of logging data is to fuse RS and RD, which are sensitive to cement and fracture response, into parameter SD. The third fusion principle of logging data is to fuse AC, CNL and DEN, which are sensitive to reservoir porosity response, into parameter ACD.

[0080] The method for fusion calculation is:

[0081] X G-S =X GRnor -X SPnor

[0082] X S-D =X RSnor +X RDnor

[0083] X A-C-D =X ACnor +X CNLnor -X DENnor

[0084] In the formula, X G-S represents the fusion value of the fusion parameter GS, X S-D represents the fusion value of the fusion parameter SD, X A-C-D represents the fusion value of the fusion parameter ACD, X GRnor represents the normalized value of GR, X SPnor represents the normalized value of SP, X RSnor represents the normalized value of RS, X RDnorrepresents the normalized value of RD, X ACnor represents the normalized value of AC, X CNLnor represents the normalized value of CNL, X DENnor Indicates the normalized value of DEN.

[0085] Step 3: cluster the fused logging data using an unsupervised machine learning algorithm, and then combine the clustered fused logging data to obtain the initial logging phase type of the entire well section;

[0086] Specifically, the K-means clustering algorithm is used to cluster each type of fused logging data into three categories. The fused logging data is one-dimensional data, so the clustering results are all high, medium, and low. The clustered fused logging data are combined. Figure 2 As shown, A1, A2, A3 represent high GS, medium GS, and low GS respectively; B1, B2, B3 represent high SD, medium SD, and low SD respectively; C1, C2, C3 represent high ACD, medium ACD, and low ACD respectively. The three fused logging data can be combined into 27 types of initial logging phases, and the codes of the initial logging phases are X1, X2, ..., X27.

[0087] Step 4: Screen the initial logging phase types based on geological constraints, and perform intersection and merging processing on the screened results to obtain the target logging phase types of the entire well section;

[0088] Specifically, the initial logging phase types are deleted based on a preset screening mechanism to obtain 16 types of initial logging phases after screening, wherein the screening mechanism is set based on geological law constraints, and the screening mechanism includes: deleting the initial logging phases containing mutually opposing diagenesis, and deleting the initial logging phases containing contradictory diagenesis and pore response characteristics.

[0089] The fusion parameter GS is sensitive to the shale content and the intensity of compaction. Generally, the higher the GS, the stronger the compaction. The fusion parameter SD is sensitive to the cementing material. Generally, the higher the SD, the stronger the cementation. The fusion parameter ACD is sensitive to the porosity. Generally, the higher the ACD, the higher the porosity. In sandstone reservoirs, diagenesis types with strong compaction and strong cementation characteristics often do not appear at the same time. Therefore, the logging phase containing high GS and high SD characteristics will not appear, that is, Figure 2 X1, X2, and X3 in the above figure; the porosity in the reservoir with strong cementation is usually not high, and it does not show the characteristics of strong cementation and medium-high porosity. Figure 2 X10, X11, X13, X19, and X20 will not appear. In addition, the logging phase types of medium-weak compaction, medium-weak cementation, and low porosity will not appear, that is, Figure 2In summary, through the screening of geological constraints based on diagenesis and pore response, the remaining logging phase types are 16. Figure 3 As shown, the codes of the screened logging phases are Y1, Y2, ..., Y16.

[0090] The GS, SD, and ACD ternary cross-plots were constructed based on the fused logging data of the 16 types of initial logging phases after screening. The ternary cross-plots were merged based on the distribution characteristics to determine the target logging phase type of the entire well section.

[0091] like Figure 4 As shown in the figure, the GS, SD, and ACD ternary cross-plots were constructed by fusion logging data of the initial logging phases after screening in the S1 well, and the six target logging phase types were determined based on the distribution characteristics of the initial logging phases after screening in the cross-plots. The codes of the target logging phase types of the S1 well are Z1, Z2, Z3, Z4, Z5, and Z6, of which Z1 only includes Y7, Z2 includes Y3, Y6, Y8, Y9, and Y12, Z3 includes Y13 and Y15, Z4 includes Y11 and Y14, Z5 includes Y2, Y5, and Y16, and Z6 includes Y1, Y4, and Y10.

[0092] Step 5: Select several sample points in the core section of the target well, mark the diagenetic facies of the sample points through cast thin section analysis, and construct a diagenetic facies data set;

[0093] 91 sample points were selected in the core section of the target well, and the cast thin sections of the 91 sample points were analyzed to determine the diagenetic parameters of the sample points. The diagenetic parameters include apparent compaction rate, cement content, surface area ratio, and fracture rate, as shown in Table 1. Based on the diagenetic parameters, the quantitative classification standard of diagenetic facies types was determined, and the reservoir of Well S1 was divided into four diagenetic facies: dense compaction phase, carbonate cementation phase, weak dissolution phase, and strong dissolution and fracture phase.

[0094] Table 1

[0095]

[0096] Step 6: Calibrate and verify the target logging facies type based on the diagenetic facies data set, obtain the diagenetic facies type corresponding to each target logging, and complete the diagenetic facies identification of the entire well section;

[0097] In this embodiment, 64 data in the diagenetic phase data set are selected to calibrate the target logging phases. The calibration results are as follows: the target logging phases corresponding to dense compaction diagenesis are Z1 and Z3, the target logging phase corresponding to carbonate cementation diagenesis is Z2, the target logging phase corresponding to weak dissolution diagenesis is Z4, and the target logging phases corresponding to strong dissolution and fracture diagenesis are Z5 and Z6. The remaining 27 data in the diagenetic phase data set are selected to verify the accuracy of the calibration results. Figure 5 As shown, 24 out of 27 test data of Well S1 were correctly identified, and the accuracy of diagenetic phase calibration results reached 88.9%.

[0098] Based on the calibration results of diagenetic facies and logging facies, the diagenetic facies identification and prediction of the entire well section of Well S1 were realized. The results are as follows: Figure 6 shown.

[0099] In order to compare the application effect of this method with the existing methods and technologies, three machine learning-based diagenetic facies logging identification methods, namely random forest, SGBoost and BP neural network, were used to carry out the same data of Well S1. The accuracy rates of the above three methods were 63.0%, 70.4% and 77.8% respectively, which were significantly lower than the accuracy rate of the method of the present invention. Therefore, this method can significantly improve the accuracy of diagenetic facies logging identification.

[0100] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A diagenetic facies identification method based on geologically constrained logging parameter fusion clustering, characterized in that: The specific steps include: Step 1: collecting logging data of the entire well section of the target well and normalizing the logging data; Step 2: Based on geological knowledge constraints, the processed logging data is fused and calculated to obtain fused logging data; Step 3: clustering the fused logging data respectively using an unsupervised machine learning algorithm, and then combining the clustered fused logging data to obtain the initial logging phase type of the entire well section; Step 4: Screening the initial logging phase type based on geological law constraints, and performing intersection and merging processing on the screened results to obtain the target logging phase type of the entire well section, wherein the geological law constraints include diagenesis and pore response characteristics; Step 5, select a number of sample points in the core section of the target well, mark the diagenetic facies of the sample points through cast thin section analysis, and construct a diagenetic facies data set; Step 6: Calibrate and verify the target logging facies type based on the diagenetic facies data set, obtain the diagenetic facies type corresponding to each target logging, and complete the diagenetic facies identification of the entire well section.

2. The diagenetic phase identification method based on geological constraints and well logging parameter fusion clustering according to claim 1 is characterized in that: The types of logging data in step 1 include: natural gamma ray GR, spontaneous potential SP, shallow resistivity RS, deep resistivity RD, acoustic wave time difference AC, compensated neutron CNL, and density DEN.

3. The diagenetic phase identification method based on geological constraints and well logging parameter fusion clustering according to claim 1 is characterized in that: The method for normalizing the logging data in step 1 is: Where, X nor represents the normalized logging data, X represents the logging data, and X min Indicates the minimum value of the logging data, X max Indicates the maximum value of the logging data.

4. The diagenetic phase identification method based on geological constraints and well logging parameter fusion clustering according to claim 1 is characterized in that: In step 2, the processed logging data is fused and calculated based on geological knowledge constraints to obtain fused logging data, which includes: Performing fusion calculation on the normalized logging data based on a preset fusion principle to obtain the fused logging data, wherein the fused logging data includes first fused logging data, second fused logging data, and third fused logging data; The fusion principle is set based on geological constraints. The fusion principle of the first fusion logging data is to fuse GR and SP, which are sensitive to mud content and lithology, into parameter GS. The fusion principle of the second fusion logging data is to fuse RS and RD, which are sensitive to cement and fracture responses, into parameter SD. The fusion principle of the third fusion logging data is to fuse AC, CNL and DEN, which are sensitive to reservoir porosity responses, into parameter ACD.

5. The diagenetic phase identification method based on geological constraints and well logging parameter fusion clustering according to claim 4 is characterized in that: The method for performing the fusion calculation is: X G-S =X GRnor -X sPnor X S-D =X RSnor +X RDnor X A-C-D =X ACnor +X CNLnor -X DENnor Where, X G-S represents the fusion value of the fusion parameter GS, X S-D represents the fusion value of the fusion parameter SD, X A-C-D represents the fusion value of the fusion parameter ACD, X GRnor represents the normalized value of GR, X SPnor represents the normalized value of SP, X RSnor represents the normalized value of RS, X RDnor represents the normalized value of RD, X ACnor represents the normalized value of AC, X CNLnor represents the normalized value of CNL, X DENnor Indicates the normalized value of DEN.

6. The diagenetic phase identification method based on geological constraints and well logging parameter fusion clustering according to claim 1 is characterized in that: In step 3, the fused logging data are clustered using an unsupervised machine learning algorithm, and then the clustered fused logging data are combined to obtain the initial logging phase type of the entire well section, including: K-means clustering algorithm was used to cluster each type of fused logging data into three categories; Based on the clustering combination results of three types of fused logging data, 27 types of initial logging facies were obtained.

7. The diagenetic phase identification method based on geological constraints and well logging parameter fusion clustering according to claim 1 is characterized in that: In step 4, the initial logging phase type is screened based on geological law constraints, and the screened results are subjected to intersection and merging processing to obtain the target logging phase type of the entire well section, including: The initial logging phase types are deleted based on a preset screening mechanism to obtain 16 types of initial logging phases after screening, wherein the screening mechanism is set based on geological law constraints, and the screening mechanism includes: deleting initial logging phases containing mutually opposing diagenesis, and deleting initial logging phases containing mutually contradictory diagenesis and pore response characteristics; Based on the fused logging data of the screened 16 types of initial logging phases, a GS, SD, and ACD ternary cross-plot is constructed, and the ternary cross-plot is merged based on the distribution characteristics to determine the target logging phase type of the entire well section.

8. The diagenetic phase identification method based on geological constraints and well logging parameter fusion clustering according to claim 1 is characterized in that: The diagenetic phase marking of the sample points by casting thin section analysis in step 5 includes: Determine the diagenetic parameters of the sample point by analyzing the casting thin section of the sample point, wherein the diagenetic parameters include apparent compaction rate, cement content, surface ratio, and crack ratio; A quantitative division standard of diagenetic phases is determined based on the diagenetic parameters, and the diagenetic phases of the sample points are marked.

9. The diagenetic phase identification method based on geological constraints and well logging parameter fusion clustering according to claim 1, characterized in that: The calibration and verification of the target logging facies type based on the diagenetic facies data set in step 6 includes: 70% of the data in the diagenetic facies data set are selected to calibrate the target logging facies type, and the remaining 30% of the data in the diagenetic facies data set are selected to verify the accuracy of the calibration result.

Images