Method for identifying shale lithofacies through three-dimensional cross plot based on logging multiple parameters
Through the three-dimensional intersection diagram method based on logging multi-parameters, the problem of difficulty in shale shale facies is solved, and more accurate and efficient shale shale facies are achieved, which is suitable for shale reservoirs with complex structures.
Patent Information
- Application Number
- CN202311509517.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-11-13
- Publication Date
- 2025-05-13
AI Technical Summary
The prior art has difficulties in identifying shale staphylococcus, especially in shale layers with heterogeneity and thin sandstone phenomena, where conventional logging curve data are difficult to accurately identify.
The method of identifying shale stag phases based on multiple parameters of well logging is adopted. By dividing the shale stag phase types in the research area, standard lithophagos sample is selected, and the applicable logging curves are screened. The optimal projection matrix is found using the Wilks criterion and Fisher discrimination method, and the distribution range of various shale stag phases is determined by the three-dimensional intersection method.
This method can more accurately identify shale staphylococcus, improve the applicability and accuracy of logging identification, especially in shale reservoirs, and has a wide application range.
Smart Images
Figure CN119986812A_ABST
Abstract
Description
Technical field:
[0001] The invention relates to the field of petroleum geological exploration and development, and in particular to a method for identifying shale lithofacies based on a three-dimensional intersection diagram of multiple logging parameters. Background technology:
[0002] There is a correlation between the lithofacies type of continental mudstone and the enrichment of shale oil. The accurate classification and identification of shale lithofacies types is an important basis for evaluating the exploration and development potential of shale gas. The logging identification of lithofacies is generally carried out through conventional logging data such as spontaneous potential (SP), density (DEN), resistivity (LLD), acoustic time difference (DT), natural gamma (GR), etc., combined with special logging data such as imaging logging and element capture logging.
[0003] Conventional logging curve data identification methods generally use the intersection plot method, radar plot method, or combine modern data methods to form a series of lithology and lithofacies logging identification methods. The original intention of selecting and using these methods is to solve the problem of lithology and lithofacies logging identification of the target layer according to the characteristics of the target reservoir section of the research block. The discrimination methods formed are highly targeted, and there are certain differences in applicable conditions and application scope. Therefore, it is of great significance to study a lithology and lithofacies logging identification method with a wide range of applications and strong applicability for the evaluation of shale oil reservoirs. The Qingshankou Formation shale layer in the Gulong area of the Songliao Basin is characterized by wide distribution, large thickness, and high organic matter abundance. There is a strong horizontal and vertical heterogeneity in the shale layer, and thin interlayer sandstone is locally developed, which brings certain difficulties to the logging identification of shale lithofacies.
[0004] Well logging identification of lithofacies is the basis for realizing well logging evaluation of unconventional reservoirs and is crucial to realizing reservoir quality evaluation based on well logging. In order to be able to identify shale lithofacies using well logging curves, it becomes very meaningful to amplify the differences in the response characteristics of logging data in various lithofacies and to integrate multiple logging curves for logging identification of shale lithofacies. Summary of the invention:
[0005] The present invention aims to solve the problem that it is difficult to identify shale lithofacies using conventional logging curve data in the background technology, and provides a method for identifying shale lithofacies using a three-dimensional intersection diagram based on logging multi-parameters. The method for identifying shale lithofacies using a three-dimensional intersection diagram based on logging multi-parameters has a wide range of applications and strong applicability, and is of great significance for the evaluation of shale oil reservoirs.
[0006] The present invention solves the problem through the following technical solution: The method for identifying shale lithofacies based on a three-dimensional intersection diagram of multiple logging parameters comprises the following steps:
[0007] Step 1: Classify the shale lithofacies types within the target interval of the study area;
[0008] Step 2: Based on the divided lithofacies type, select standard lithofacies samples; based on the selected standard lithofacies samples, according to the logging response characteristics of shale lithofacies in the target layer of the study area, preliminarily select logging parameters that are relatively sensitive to shale lithofacies; use the Wilks criterion to screen the logging curves suitable for shale lithofacies division;
[0009] Step 3: Based on the screened logging curves, use the Fisher discriminant method to find an optimal projection matrix so that the same lithofacies have the smallest intra-class discreteness and different lithofacies have the largest inter-class discreteness. Use the three-dimensional intersection method to determine the distribution range of each type of shale lithofacies in three-dimensional space and perform logging identification on the shale lithofacies.
[0010] Furthermore, the method for dividing the lithofacies types in the target layer section of the study area in step S1 is:
[0011] Core observation was carried out on the cored well section of the target layer in the study area, and the lithofacies types in the target layer in the study area were divided according to the classification principles of organic matter abundance, sedimentary structure and rock type that control shale oil accumulation by means of comprehensive thin section identification, element analysis and mercury injection test.
[0012] Furthermore, the shale lithofacies in the target interval of the study area is classified as L type.
[0013] Furthermore, the method for screening the well logging curves suitable for shale lithofacies classification in step 2 comprises the following steps:
[0014] 1.1. Based on the selected standard lithofacies samples and the logging response characteristics of shale lithofacies in the target layer of the study area, g logging parameters that are relatively sensitive to shale lithofacies are preliminarily selected; the selected g logging parameters that are relatively sensitive to shale lithofacies are extremeized;
[0015] 1.2. Use the Wilks criterion to screen the preliminarily selected logging curves and select the logging parameters that can be used as shale lithofacies discrimination parameters.
[0016] Furthermore, the step 1.1 preliminarily selects g logging parameters that are sensitive to shale lithofacies, including compensated neutron (CNL), density (DEN), resistivity (LLD), sonic time difference (DT), microsphere (MSFL), dry weight of calcium (DWGA), dry weight of magnesium (DWMG), dry weight of aluminum (DWAL) and dry weight of silicon (DWSI).
[0017] Furthermore, in step 1.1, the g logging parameters initially selected and sensitive to shale lithofacies are maximized, and the expression used is:
[0018]
[0019] Where: vj v i The range conversion value of the jth sampling point of the logging curve; N is the number of sampling points in the selected shale layer; i = 1, 2, ..., g;
[0020] Among them: for the initially selected logging parameters that are sensitive to shale lithofacies, the following are set: CNL range conversion mean v1, DEN range conversion mean v2, LLD range conversion mean v3, DT range conversion mean v4, MSFL range conversion mean v5, DWGA extreme value conversion mean v6, DWMG extreme value conversion mean v7, GR range conversion mean v8, LLS range conversion mean v9.
[0021] Furthermore, the step 1.2 uses the Wilks criterion to screen the well logging curves and select the method that can be used as the shale lithofacies logging identification parameter, including:
[0022] The g logging extreme value parameters that are relatively sensitive to the preliminary selected shale lithofacies are screened, the variables that are irrelevant or have little influence on the discriminant function are eliminated, a reasonable discriminant equation is constructed, and finally the logging discriminant parameters of the shale lithofacies are extracted.
[0023] Furthermore, the specific steps of constructing a reasonable discriminant equation are:
[0024] First, samples are taken from P populations b1, b2, ..., b p Take out n1, n2, ..., n respectively p Samples, each sample has g logging parameters, and the observation samples constitute a g-dimensional matrix, then its component variables are:
[0025]
[0026] Where: is the g-th logging curve data of the k-th sample of the P-th lithofacies type;
[0027] p=1,2,…,P;k=1,2,…,n p ;
[0028] It is known from experience that the cumulative probability of each variable conforms to the normal distribution, so the intra-class deviation matrix is established as:
[0029]
[0030] The total deviation matrix is established as:
[0031]
[0032] The statistic U is:
[0033] U=|W| / |T| (5)
[0034] Among the h variables that have been introduced, test the variable x (r) Should it be removed, then the statistic U r At this time Its discriminant equation is:
[0035]
[0036] If the F value is less than the critical value F' of the test level a , then x (r) If the parameters are not met, they are eliminated, otherwise they are retained, and finally m parameters are extracted as the logging discrimination parameters of shale lithofacies.
[0037] Furthermore, the step 3 uses the Fisher discriminant method to find an optimal projection matrix so that the same type of lithofacies has the smallest intra-class discreteness and different types of lithofacies have the largest inter-class discreteness. The specific method is:
[0038] After extracting m parameters from each of the obtained P samples and maximizing them, the sample mean u of sample class i is i for:
[0039]
[0040] The overall sample mean u is:
[0041]
[0042] The inter-class scatter matrix S b for:
[0043]
[0044] The intra-class scatter matrix S w for:
[0045]
[0046] The shale reservoirs are divided into l categories, that is, l populations. According to multivariate statistical theory, the Fisher discriminant function is used:
[0047]
[0048] Where: is the optimal projection vector that maximizes the objective function;
[0049] Substituting formula (9) and formula (10) into formula (11), we get:
[0050]
[0051] The solved eigenvalues and eigenvectors are combined into a typical discriminant function equation matrix to discriminate the samples to be judged.
[0052] Furthermore, the step S3 uses a three-dimensional intersection method to determine the distribution range of various shale lithofacies in three-dimensional space, and the method for logging identification of shale lithofacies includes the following steps:
[0053] 2.1) Based on the Fisher discriminant method, the solved eigenvalues and eigenvectors are combined into a function matrix, and the eigenvectors corresponding to the first three largest eigenvalues of the matrix are taken to construct the normalized eigenvectors α1, α2, and α3;
[0054] 2.2) Project the sample data of type l lithofacies to be identified on α1, α2, and α3, where X is the projection direction of α1, Y is the projection direction of α2, and Z is the projection direction of α3. Each type of sample has its own cluster center. The sample points are distributed in three-dimensional space, and their respective distribution spaces are divided to perform logging identification on shale lithofacies.
[0055] Compared with the above background technology, the present invention has the following beneficial effects:
[0056] In practical applications, the method for identifying shale lithofacies based on the three-dimensional intersection diagram of logging multi-parameters of the present invention can comprehensively consider multiple logging curves as parameters, thereby providing a method that is better adapted to shale reservoir lithofacies logging identification. The method has the characteristics of stable algorithm, large amount of information, high accuracy, convenient application, and fast and intuitive. It can be better applied to shale reservoirs where the logging curve has a poor reflection effect on shale lithofacies and conventional two-dimensional intersection diagram methods and other methods cannot accurately distinguish various lithofacies, and can play a huge role in finding oil and gas resources and reservoir evaluation. Description of the drawings:
[0057] Attached Figure 1 This is a flow chart of a method for identifying shale lithofacies based on a three-dimensional intersection diagram of multiple logging parameters of the present invention;
[0058] Attached Figure 2 It is a three-dimensional spatial distribution diagram of various shale facies in an embodiment of the present invention;
[0059] Attached Figure 3 This is a shale lithofacies logging interpretation diagram according to an embodiment of the present invention. Specific implementation method:
[0060] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the protection scope of the present invention.
[0061] like Figure 1 As shown, a method for identifying shale lithofacies based on a three-dimensional intersection diagram of multiple logging parameters includes the following steps:
[0062] Step 1: Divide the lithofacies types within the target interval of the study area;
[0063] Core observation was carried out on the cored well section of the target layer in the study area, and the lithofacies types in the target layer in the study area were divided according to the classification principles of organic matter abundance, sedimentary structure and rock type that control shale oil accumulation by means of comprehensive thin section identification, element analysis and mercury injection test.
[0064] The shale lithofacies in the target layer of the study area is divided into L type.
[0065] Step 2: Based on the divided lithofacies type, select standard lithofacies samples; based on the selected standard lithofacies samples, according to the logging response characteristics of shale lithofacies in the target layer of the study area, preliminarily select logging parameters that are relatively sensitive to shale lithofacies; use the Wilks criterion to screen the logging curves suitable for shale lithofacies division; specifically include the following steps:
[0066] 1.1. Based on the selected standard lithofacies samples and the logging response characteristics of shale lithofacies in the target layer of the study area, g logging parameters that are relatively sensitive to shale lithofacies are preliminarily selected; the selected g logging parameters that are relatively sensitive to shale lithofacies are extremeized;
[0067] The g logging parameters that are sensitive to shale lithofacies are initially selected, including compensated neutron (CNL), density (DEN), resistivity (LLD), sonic time difference (DT), microsphere (MSFL), dry weight of calcium (DWGA), dry weight of magnesium (DWMG), dry weight of aluminum (DWAL) and dry weight of silicon (DWSI).
[0068] The g logging parameters that are initially selected and sensitive to shale lithofacies are maximized, and the expression used is:
[0069]
[0070] Where: v j v iThe range conversion value of the jth sampling point of the logging curve; N is the number of sampling points in the selected shale layer; i = 1, 2, ..., g;
[0071] Among them: for the initially selected logging parameters that are sensitive to shale lithofacies, the following are set: CNL range conversion mean v1, DEN range conversion mean v2, LLD range conversion mean v3, DT range conversion mean v4, MSFL range conversion mean v5, DWGA extreme value conversion mean v6, DWMG extreme value conversion mean v7, GR range conversion mean v8, LLS range conversion mean v9.
[0072] 1.2. Using the Wilks criterion, screen the preliminarily selected well logging curves and select the well logging parameters that can be used as shale lithofacies identification parameters; the specific methods include:
[0073] The g logging extreme value parameters that are relatively sensitive to the preliminary selected shale lithofacies are screened, the variables that are irrelevant or have little influence on the discriminant function are eliminated, a reasonable discriminant equation is constructed, and finally the logging discriminant parameters of the shale lithofacies are extracted.
[0074] The specific steps of constructing a reasonable discriminant equation are:
[0075] First, samples are taken from P populations b1, b2, ..., b p Take out n1, n2, ..., n respectively p Samples, each sample has g logging parameters, and the observation samples constitute a g-dimensional matrix, then its component variables are:
[0076]
[0077] Where: is the number of the g-th logging curve of the k-th sample of the P-th lithofacies type;
[0078] p=1,2,…,P;k=1,2,…,n p ;
[0079] It is known from experience that the cumulative probability of each variable conforms to the normal distribution, so the intra-class deviation matrix is established as:
[0080]
[0081] The total deviation matrix is established as:
[0082]
[0083] The statistic U is:
[0084] U=|W| / |T| (5)
[0085] Among the h variables that have been introduced, test the variable x(r) Should it be removed, then the statistic U r At this time Its discriminant equation is:
[0086]
[0087] If the F value is less than the critical value F' of the test level a , then x (r) If the parameters are not met, they are eliminated, otherwise they are retained, and finally m parameters are extracted as the logging discrimination parameters of shale lithofacies.
[0088] Step 3: Based on the screened logging curves, use the Fisher discriminant method to find an optimal projection matrix so that the same lithofacies have the smallest intra-class discreteness and different lithofacies have the largest inter-class discreteness. Use the three-dimensional intersection method to determine the distribution range of each type of shale lithofacies in three-dimensional space and perform logging identification on the shale lithofacies.
[0089] The specific method of finding an optimal projection matrix based on the screened logging curves and using the Fisher discriminant method so that the same type of lithofacies has the smallest intra-class discreteness and different types of lithofacies have the largest inter-class discreteness is as follows:
[0090] After extracting m parameters from each of the obtained P samples and maximizing them, the sample mean u of sample class i is i for:
[0091]
[0092] The overall sample mean u is:
[0093]
[0094] The inter-class divergence matrix S b for:
[0095]
[0096] The intra-class scatter matrix S w for:
[0097]
[0098] The shale reservoirs are divided into l categories, that is, l populations. According to multivariate statistical theory, the Fisher discriminant function is used:
[0099]
[0100] Where: is the optimal projection vector that maximizes the objective function;
[0101] Substituting formula (9) and formula (10) into formula (11), we get:
[0102]
[0103] The solved eigenvalues and eigenvectors are combined into a typical discriminant function equation matrix to discriminate the samples to be judged.
[0104] The method of using the three-dimensional intersection method to determine the distribution range of various shale lithofacies in three-dimensional space and to perform logging identification on shale lithofacies includes the following steps:
[0105] 2.1) Based on the Fisher discriminant method, the solved eigenvalues and eigenvectors are combined into a function matrix, and the eigenvectors corresponding to the first three largest eigenvalues of the matrix are taken to construct the normalized eigenvectors α1, α2, and α3;
[0106] 2.2) Project the sample data of type l lithofacies to be identified on α1, α2, and α3, where X is the projection direction of α1, Y is the projection direction of α2, and Z is the projection direction of α3. Each type of sample has its own cluster center. The sample points are distributed in three-dimensional space, and their respective distribution spaces are divided to perform logging identification on shale lithofacies.
[0107] Example 1
[0108] By accurately identifying the shale lithofacies in the 2548-2570m section of Well X in the Gulong area of the Songliao Basin, the method for identifying shale lithofacies based on a three-dimensional intersection diagram of multiple logging parameters of the present invention is specifically described, including the following steps:
[0109] (1) As the organic matter abundance and physical properties of the shale in the target layer of the study area have a decisive influence on the formation of shale oil reservoirs, a comprehensive naming method combining the sedimentary structure that controls the physical properties and the organic matter abundance is adopted in the classification and naming of shale lithofacies. Shale lithofacies are divided into five categories: high-organic layered shale facies, high-organic laminated shale facies, medium-organic laminated shale facies, medium-organic massive dolomite facies, and low-organic shale facies.
[0110] (2) After depth correction, environmental correction, and core observation comparison of the well logging curves of the target layer in the study area, 9 logging parameters were selected, including compensated neutron (CNL), density (DEN), resistivity (LLD), sonic time difference (DT), microsphere (MSFL), dry weight of calcium (DWGA), dry weight of magnesium (DWMG), dry weight of aluminum (DWAL), and dry weight of silicon (DWSI), according to the morphology of shale lithofacies curves and logging response characteristics.
[0111] (3) Use the extreme value transformation formula and the extreme value transformation deviation square sum formula to perform extreme value transformation on the selected curve. The calculation formula is as follows:
[0112]
[0113] Where: v j v i The range conversion value of the jth sampling point of the logging parameter; N is the number of sampling points in the selected shale layer, CNL range conversion mean v1, DEN range conversion mean v2, LLD range conversion mean v3, DT range conversion mean v4, MSFL range conversion mean v5, DWGA extreme value conversion mean v6, DWMG extreme value conversion mean v7, DWAL range conversion mean v8, DWSI range conversion mean v9.
[0114] (4) Using the Wilks criterion, the preliminarily selected logging parameters that can characterize the shale lithofacies category are screened. By reducing the dimensionality of the multidimensional logging data, the variables that are irrelevant or have little influence on the discriminant function are eliminated, and a reasonable discriminant equation is constructed. Finally, m parameters are selected as the evaluation parameters of the logging lithofacies. This method can achieve the effect of dimensionality reduction for subsequent calculations.
[0115] First, select the selected samples from P populations b1, b2, ..., b p Take out n1, n2, ..., n respectively p Samples, each sample has m logging parameters, and the sample constitutes an m-dimensional matrix of observation samples, then its component variables are:
[0116]
[0117] Where: is the g-th logging parameter of the P-th sample, p=1,2,…,P; k=1,2,…,n p ;
[0118] It is known from experience that the cumulative probability of each variable conforms to the normal distribution, so the intra-class deviation matrix is established as:
[0119]
[0120] The total deviation matrix is established as:
[0121]
[0122] The statistic U is:
[0123] U=|W| / |T|
[0124] Among the h variables that have been introduced, test the variable x (r) Should it be removed, then the statistic U r At this time Its discriminant equation is:
[0125]
[0126] If the F value is less than the critical value F' of the test level a , then x (r) Remove, otherwise keep.
[0127] Using this method, the above 9 parameters v1, v2, v3, v4, v5, v6, v7, v8, v9 were tested, and finally v1, v2, v3, v4, v5, v6, v77 parameters were selected as logging parameters for shale lithofacies identification. The logging parameter contribution rate is shown in Table 1; the logging response characteristics of different shale lithofacies are shown in Table 2.
[0128] Table 1 Contribution rate of logging parameters
[0129]
[0130]
[0131] Table 2 Logging response characteristics of different shale lithofacies
[0132]
[0133] (5) Establishment of discriminant function.
[0134] After extracting m parameters from each sample and maximizing them, the sample mean u of sample class i i for:
[0135]
[0136] The overall sample mean u is:
[0137]
[0138] The inter-class scatter matrix S b for:
[0139]
[0140] The intra-class scatter matrix S w for:
[0141]
[0142] The shale reservoirs are divided into classes, i.e., l populations. According to multivariate statistical theory, the Fisher discriminant function is used:
[0143]
[0144] Where: is the optimal projection vector that maximizes the objective function.
[0145] S b and S w Bring in The Fisher discriminant function obtained is:
[0146]
[0147] The solved eigenvalues and eigenvectors are combined into a typical discriminant function equation matrix to discriminate the samples to be judged.
[0148] (6) Three-dimensional spatial distribution of shale facies
[0149] Based on the above discriminant function, the eigenvectors corresponding to the first three largest eigenvalues of the matrix are taken to construct unitized eigenvectors α1, α2 and α3. The weight functions corresponding to α1, α2 and α3 can assign different function eigenvalues according to the contribution rate of the seven logging parameters to the shale lithofacies logging discrimination. That is:
[0150] a1=-0.40*CNL-0.051*MSFL+17.836*DWCA+0.276*
[0151] DWMG+0.196*LLD-0.232*DT+1.619*DEN-10.343
[0152] a2=-0.053*CNL-0.033*MSFL+28.133*DWCA-0.094*
[0153] DWMG+0.157*LLD-40.044*DT+6.151*DEN-6.534
[0154] a3=-0.035*CNL-0.003*MSFL+34.291*DWCA-0.006*
[0155] DWMG+0.062*LLD-0.059*DT-1.531*DEN-15.748
[0156] Figure 2 It is a three-dimensional spatial distribution diagram of various shale facies. Figure 2 As shown in the figure, the five types of lithofacies samples to be identified are projected onto vectors α1, α2 and α3. The five different shale lithofacies are well distinguished, and each type of lithofacies has its own cluster center. Due to the effect of three-dimensional space displayed on the plane, the display effect will be different at different observation angles. In practical applications, the distribution of various lithofacies is relatively clear and definite. Its distribution range:
[0157] 1) High organic matter layered shale phase a1 distribution range 2 to 4; a2 distribution range -2 to 4; a3 distribution range: -3 to 2
[0158] 2) High organic matter laminated shale phase a1 distribution range 0~2; a2 distribution range -1.9~2.3; a3 distribution range: -0.8~2
[0159] 3) Medium organic laminated shale phase a1 distribution range -2 to 1; a2 distribution range -4 to 0; a3 distribution range: -2.5 to 0
[0160] 4) Medium organic massive dolomite phase a1 distribution range -3 ~ 0; a2 distribution range -2 ~ 2; a3 distribution range: 0 ~ 3
[0161] 5) The distribution range of low organic matter shale phase a1 is -2 to 2; the distribution range of a2 is 0 to 4; the distribution range of a3 is: -1 to 4.
[0162] (7) Inspection of shale lithofacies identification results by well logging: The inspection method is to use the well logging identification conclusions of the lithofacies to conduct a comparative inspection of the lithofacies of the entire well section of the target layer with the well logging identification conclusions and core analysis.
[0163] Figure 3 It is a comprehensive identification curve of various lithofacies in the study area in the embodiment of the present invention. Lithofacies logging identification was performed on the 2548-2570m section of the non-modeling well X. The accuracy rate was 93.54%. The results showed that except for two low-organic shale thin layers that were not identified and displayed as medium-organic laminar shale phases, the remaining layers were correctly identified with an accuracy rate of 93.54%. By back-judging the lithofacies logging identification effect of the coring wells, the comparison of the 76m coring sections of the three wells was statistically analyzed, and the lithofacies of 121 layers were correctly identified, with a compliance rate of up to 90.3%. It shows that the established model can be well applied to the identification of shale lithofacies in the study area.
Claims
1. A method for identifying shale lithofacies based on a three-dimensional intersection diagram of multiple logging parameters, characterized in that: The following steps are involved: Step 1: Classify the shale lithofacies types within the target interval of the study area; Step 2: Based on the classified lithofacies types, select standard lithofacies samples; based on the selected standard lithofacies samples, according to the logging response characteristics of shale lithofacies in the target layer of the study area, preliminarily select logging parameters that are relatively sensitive to shale lithofacies; Using Wilks criterion, the well logging curves suitable for shale lithofacies division are screened; Step 3: Based on the screened logging curves, use the Fisher discriminant method to find an optimal projection matrix so that the same lithofacies have the smallest intra-class discreteness and different lithofacies have the largest inter-class discreteness. Use the three-dimensional intersection method to determine the distribution range of each type of shale lithofacies in three-dimensional space and perform logging identification on the shale lithofacies.
2. The method for identifying shale lithofacies based on a three-dimensional intersection diagram of well logging multi-parameters according to claim 1 is characterized in that: Step S1: The method for dividing the lithofacies types in the target layer of the study area is: Core observation was carried out on the cored well section of the target layer in the study area, and the shale lithofacies types in the target layer in the study area were divided according to the classification principles of organic matter abundance, sedimentary structure and rock type that control shale oil accumulation by means of comprehensive thin section identification, element analysis and mercury injection test.
3. The method for identifying shale lithofacies based on a three-dimensional intersection diagram of well logging multi-parameters according to claim 2 is characterized in that: The lithofacies in the target layer of the study area are divided into L type.
4. The method for identifying shale lithofacies based on a three-dimensional intersection diagram of well logging multi-parameters according to claim 3 is characterized by: The method of step 2 for screening the well logging curves suitable for shale lithofacies division, The following steps are involved: 1.
1. Based on the selected standard lithofacies samples and the logging response characteristics of the shale lithofacies in the target layer of the study area, g logging curves that are relatively sensitive to the shale lithofacies are preliminarily selected; the selected g logging curves that are relatively sensitive to the shale lithofacies are extremeized; 1.
2. Use the Wilks criterion to screen the initially selected logging curves; Select logging parameters that can be used as shale lithofacies identification parameters.
5. The method for identifying shale lithofacies based on a three-dimensional intersection diagram of well logging multi-parameters according to claim 4 is characterized in that: The step 1.1 preliminarily selects g logging parameters that are relatively sensitive to shale lithofacies, including: compensated neutron, density, resistivity, acoustic wave time difference, microsphere, dry weight of calcium, dry weight of magnesium, dry weight of aluminum and dry weight of silicon.
6. The method for identifying shale lithofacies based on a three-dimensional intersection diagram of well logging multi-parameters according to claim 5, characterized in that: In step 1.1, the g logging curves initially selected to be sensitive to shale lithofacies are extremeized, and the expression used is: Where: v j v i The range conversion value of the jth sampling point of the logging curve; N is the number of sampling points in the selected shale layer; i = 1, 2, ..., g; Among them: for the initially selected logging parameters that are sensitive to shale lithofacies, the following are set: CNL range conversion mean v1, DEN range conversion mean v2, LLD range conversion mean v3, DT range conversion mean v4, MSFL range conversion mean v5, DWGA extreme value conversion mean v6, DWMG extreme value conversion mean v7, GR range conversion mean v8, LLS range conversion mean v9.
7. The method for identifying shale lithofacies based on a three-dimensional intersection diagram of well logging multi-parameters according to claim 6, characterized in that: The step 1.2 uses the Wilks criterion to screen the preliminarily selected logging curves; the method for selecting the logging discriminant parameters that can be used as shale lithofacies is as follows: g logging extreme value parameters that are relatively sensitive to the preliminarily selected shale lithofacies are screened, variables that are irrelevant to the discriminant function or have little influence are eliminated, a reasonable discriminant equation is constructed, and finally the logging discriminant parameters as shale lithofacies are extracted.
8. The method for identifying shale lithofacies based on a three-dimensional intersection diagram of well logging multi-parameters according to claim 7, characterized in that: Use the Wilks criterion to screen the initially selected well logging curves; Select logging parameters that can be used as shale lithofacies identification. The specific steps include: First, samples are taken from P populations b1, b2, ..., b p Take out n1, n2, ω, n respectively p Samples, each sample has g logging parameters, and the observation samples constitute a g-dimensional matrix, then its component variables are: Where: is the g-th logging curve data of the k-th sample of the P-th lithofacies type; p=1,2,ω,P;k=1,2,…,n p ; It is known from experience that the cumulative probability of each variable conforms to the normal distribution, so the intra-class deviation matrix is established as: The total deviation matrix is established as: The statistic U is: U=|W| / |T| (5) Among the h variables that have been introduced, test the variable x (r) Should it be removed, then the statistic U r At this time Its discriminant equation is: If the F value is less than the critical value F' of the test level a , then x (r) If the parameters are not met, they are eliminated, otherwise they are retained, and finally m parameters are extracted as the logging discrimination parameters of shale lithofacies.
9. The method for identifying shale lithofacies based on a three-dimensional intersection diagram of well logging multi-parameters according to claim 8, characterized in that: The step 3 uses the Fisher discriminant method to find an optimal projection matrix so that the same type of lithofacies has the smallest intra-class discreteness and different types of lithofacies have the largest inter-class discreteness. The specific method is as follows: After extracting m parameters from each of the obtained P samples and maximizing them, the sample mean u of sample class i is i for: The overall sample mean u is: The inter-class scatter matrix S b for: The intra-class scatter matrix S w for: The shale reservoirs are divided into l categories, that is, l populations. According to multivariate statistical theory, the Fisher discriminant function is used: Where: is the optimal projection vector that maximizes the objective function; Substituting formula (9) and formula (10) into formula (11), the obtained Fisher discriminant function is: Using Fisher discriminant function, the solved eigenvalues and eigenvectors are combined into a typical discriminant function equation matrix.
10. The method for identifying shale lithofacies based on a three-dimensional intersection diagram of multiple logging parameters according to claim 9, characterized in that: Step S3 uses a three-dimensional intersection method to determine the distribution range of various shale lithofacies in three-dimensional space. The method for logging identification of shale lithofacies includes the following steps: 2.1) Based on the Fisher discriminant method, the function matrix composed of the solved eigenvalues and eigenvectors is taken, and the eigenvectors corresponding to the first three largest eigenvalues of the matrix are used to construct the normalized eigenvectors α1, α2, and α3; 2.2) Project the sample data of type l lithofacies to be identified on α1, α2, and α3, where X is the projection direction of α1, Y is the projection direction of α2, and Z is the projection direction of α3; each type of sample has its own cluster center, and the sample points are distributed in three-dimensional space, dividing their respective distribution spaces, and performing logging identification on shale lithofacies.