A method for calculating water coning critical production of fractured-vug carbonate reservoir with bottom water and application thereof

Abstract

The application discloses a kind of carbonate fractured-vug type bottom water reservoir water cone critical production calculation method and application thereof.The method of the application includes the comprehensive use of fuzzy clustering analysis method and BP artificial neural network method to calculate the water cone critical production Q 临界产量 Of carbonate fractured-vug type bottom water reservoir, overcomes the difficulty of parameter selection of existing water cone critical production formula, improves the prediction accuracy of water cone critical production, and can more accurately predict the water cone critical production of carbonate fractured-vug type bottom water reservoir.

Description

Technical Field

[0001] This invention belongs to the field of oil and gas reservoir energy development technology, specifically relating to a method for calculating the critical production of water cone in carbonate fractured-vuggy bottom water reservoirs and its application. Background Technology

[0002] Accurately determining the critical production rate of the water cone is crucial for the efficient development of bottom-water reservoirs. Many scholars, both domestically and internationally, have established corresponding critical production rate prediction formulas based on different physical models. These mainly include: the Dupuit formula, the modified Dupuit formula, the Chaperon formula, the Schols formula, the Hoyland, Papatzacos & Skjaeveland (HPS) formula, the Chierici formula, the Craft & Hawkins formula, the baffle theory formula, the Li Chuanliang formula, the Cheng Linsong formula, and the Dou Hongen formula, among others. However, on the one hand, these formulas are established specifically for bottom-water reservoirs, limiting their applicability; on the other hand, some parameters in these formulas are difficult to determine in the early stages of well production, and even in the later stages of oilfield development, some parameters remain difficult to accurately determine, especially for carbonate fracture-vuggy bottom-water reservoirs, where the reservoir heterogeneity is strong, making parameter value determination very difficult. Therefore, it is usually necessary to assume some parameter values ​​and roughly estimate the critical production rate of the well, which leads to a significant reduction in the accuracy of critical production rate calculations. Summary of the Invention

[0003] In view of the problems existing in the prior art, the purpose of this invention is to provide a method for calculating the critical production rate of water cone in carbonate fractured-vuggy bottom water reservoirs. This method can accurately predict the critical production rate of water cone in carbonate fractured-vuggy bottom water reservoirs, providing a strong basis for predicting bottom water cone risk, effectively guiding the adjustment of oil well measures, and improving development results.

[0004] Therefore, in a first aspect, the present invention provides a method for calculating the critical water cone production of carbonate fractured-vuggy bottom-water reservoirs. The calculation method includes: comprehensively utilizing fuzzy clustering analysis and BP artificial neural network methods to calculate the critical water cone production Q of carbonate fractured-vuggy bottom-water reservoirs. 临界产量 , where Q 临界产量 =a×Q 模糊聚类 +(1-a)×Q 神经网络 , a∈[0,1], Q 模糊聚类 Q is the critical yield obtained based on fuzzy clustering analysis. 神经网络 The critical yield obtained according to the BP artificial neural network method is determined by performing regression analysis and practical verification on the critical yield values ​​of sample wells with known critical yields and the critical yield values ​​of the sample wells calculated by fuzzy clustering analysis and BP artificial neural network method, respectively.

[0005] According to the present invention, the regression analysis and practical verification can be performed, for example, by performing regression analysis on the critical production values ​​of several (at least two) sample wells with known critical production rates and the critical production values ​​of these sample wells calculated by fuzzy clustering analysis and BP artificial neural network method respectively, or further combining the experience of those skilled in the art and expert judgment to determine the α value, or further verifying the determined α value and the critical production rate calculation formula with several known critical production rate sample wells to obtain the α value that is closest to the predicted result and the actual result.

[0006] According to a preferred embodiment of the present invention, when a = 0.43, the critical water cone production Q of the carbonate fracture-vuggy bottom water reservoir is determined. 临界产量 =0.43×Q 模糊聚类 +0.57×Q 神经网络 .

[0007] According to some embodiments of the present invention, the fuzzy clustering analysis method includes the following steps:

[0008] (1) Standardization of raw data of sample wells: wells that have been put into production and development in carbonate fractured-vuggy bottom water reservoirs are selected as sample wells, and the analysis indicators of the sample wells are selected to obtain the original matrix. The raw data of the analysis indicators are normalized to the closed interval [0, 1] to obtain the standardized data of the raw data of each analysis indicator of the sample wells.

[0009] (2) Constructing a similarity matrix: Based on the obtained standardized data, calculate the similarity coefficients between each index of each sample well and construct a fuzzy similarity matrix;

[0010] (3) Fuzzy clustering of sample wells: Based on the fuzzy similarity matrix in step (2), the sample wells are classified using MATLAB software;

[0011] (4) Classification of target wells and prediction of critical production: Based on the data of the target wells corresponding to the analysis indicators, the similarity coefficient between the target wells and the sample wells is calculated. The sample wells with the highest similarity to the target wells are obtained, and the target wells are classified into the sample well category. The average critical production of the water cone of the sample wells in this category is the critical production of the water cone of the target well, Q. 模糊聚类 .

[0012] According to some embodiments of the present invention, the BP artificial neural network method includes the following steps:

[0013] 1) Determine the input and output indicators of the BP artificial neural network: the input indicators include the reservoir characteristic indicators of the sample wells and the production dynamic indicators of the sample wells, and the output indicator is the critical production rate of the oil well;

[0014] 2) Input index digitization: The input indicators of the sample wells are digitized to obtain the input vector of the sample wells:

[0015] 3) Optimization of BP artificial neural network model structure: By adjusting the number of hidden layers and the number of nodes in each layer, the structure of the BP artificial neural network model is optimized by using trial and error comparison to determine the number of hidden layers and the number of nodes in each layer;

[0016] 4) Training the BP artificial neural network model: Divide the sample well into n parts, one part is the test sample and the remaining n-1 parts are the training samples. Normalize the input and output data, give the weights and thresholds of each layer, train the samples and adjust the corresponding parameters until the model meets the error requirements, and obtain the BP artificial neural network model.

[0017] 5) Prediction of the critical water cone production rate of the target well: Based on the trained BP artificial neural network model, the corresponding input index of the target well is input to obtain the predicted critical water cone production rate Q of the target well. 神经网络 .

[0018] According to some embodiments of the present invention, the original matrix is: Where N represents the number of sample wells, and K represents the number of analytical indicators for each sample well.

[0019] According to some embodiments of the present invention, N is not less than 10.

[0020] According to some embodiments of the present invention, the analytical indicators include: reservoir permeability, crude oil viscosity, crude oil density, crude oil volume coefficient, oil-water density difference, reservoir thickness, perforation thickness, average production, and oil pressure.

[0021] According to some embodiments of the present invention, the raw data standardization method includes:

[0022] a. For the "proliferation" indicator, the data standardization method of Formula I is adopted as follows:

[0023]

[0024] b. For the "minimizing" indicator, the data standardization method of Equation II is adopted as follows:

[0025]

[0026] Among them, X ij X′ represents the standardized value of the original data of well j in sample i. ij X represents the actual value of the original data of the index of sample well j in sample i. jmin Represents the minimum value of the j-th index parameter, X jmax This represents the maximum value among the j-th index parameters.

[0027] According to the present invention, the "larger" index refers to an index that is as large as possible.

[0028] According to the present invention, the "smaller" index refers to an index that is as small as possible.

[0029] According to some embodiments of the present invention, in step (2), the similarity coefficient is calculated using the cosine of the included angle method. Where, r ij x represents the similarity coefficient between sample well i and oil well j; ik ' represents the normalized value of index k for sample well i; x jk ' represents the normalized value of the index k for sample well j. If r ii =0, indicating that the two wells are completely unrelated; if r = 0, it means that the two wells are completely unrelated. ij =1 indicates that the two wells are completely similar or identical.

[0030] According to some embodiments of the present invention, in step (3), the maximum tree method is used to classify the sample wells by inputting a horizontal threshold λ∈[0,1].

[0031] According to some embodiments of the present invention, in step 1) of the BP artificial neural network method, the reservoir characteristic indicators of the sample well include reservoir type, interlayer conditions, formation energy conditions, oil-water distribution pattern and bottom water communication mode.

[0032] According to some embodiments of the present invention, in step 1) of the BP artificial neural network method, the production dynamic indicators of the sample well include production dynamic indicators during the waterless period, production dynamic indicators during the low water-cut period, production dynamic indicators during the medium water-cut period, and production dynamic indicators during the high water-cut period. Preferably, the production dynamic indicators of the sample well include production time during the waterless period, average liquid production during the waterless period, production time during the low water-cut period, average liquid production during the low water-cut period, and average water cut during the low water-cut period; production time during the medium water-cut period, average liquid production during the medium water-cut period, and average water cut during the medium water-cut period; and production time during the high water-cut period, average liquid production during the high water-cut period, and average water cut during the high water-cut period.

[0033] According to some embodiments of the present invention, in step 2) of the BP artificial neural network method, the reservoir characteristic indicators are digitized, which is performed by a method including the following steps:

[0034] a) The different impacts of the reservoir characteristic indicators on the critical production rate of the water cone are divided into at least five evaluation levels: good, relatively good, moderate, relatively poor, and poor. Each evaluation level corresponds to an evaluation vector consisting of 5 numbers, the sum of which is 1. The evaluation vector is determined by combining the experience of those skilled in the art and expert scoring. Preferably, the evaluation vector is: {Good: 0.83, 0.16, 0.01, 0.00, 0.00; Relatively good: 0.16, 0.68, 0.14, 0.02, 0.00; Moderate: 0.02, 0.27, 0.42, 0.27, 0.02; Relatively poor: 0.00, 0.02, 0.14, 0.68, 0.16; Poor: 0.00, 0.00, 0.01, 0.16, 0.83}.

[0035] b) Assign different weights to each reservoir characteristic index to obtain the reservoir characteristic index weight vector.

[0036] c) Determine the evaluation level corresponding to the reservoir characteristic index of the oil well, and thus obtain the evaluation vector matrix corresponding to the oil well;

[0037] d) Multiply the evaluation vector matrix with the weight vector to obtain the digital results of the reservoir characteristic indicators of the target well.

[0038] According to some embodiments of the present invention, the evaluation vector in step a) is shown in Table 1 below:

[0039] Table 1. Evaluation Set of Indicator Levels

[0040]

[0041] According to some embodiments of the present invention, in step c), the evaluation level corresponding to the corresponding reservoir characteristic index of the target oil well can be determined by dividing each reservoir characteristic index into 5 factor levels according to their different effects on the critical production of the water cone. The 5 factor levels correspond to the levels of good, relatively good, medium, poor, and bad, respectively.

[0042] According to some embodiments of the present invention, the reservoir characteristic indicators include reservoir type, formation energy conditions, interlayer conditions, bottom water connection method, and the location of the production section.

[0043] According to some embodiments of the present invention, the factor levels of the fissure-cavity reservoir type include: cave type, small cave type, fissure-cavity, small fissure type, and fissure type, which correspond to good, relatively good, medium, relatively poor, and poor levels, respectively.

[0044] According to some embodiments of the present invention, the factor levels of the formation energy conditions include: weak, relatively weak, moderate, relatively strong, and strong, which correspond to good, relatively good, moderate, relatively poor, and poor levels, respectively.

[0045] According to the present invention, the formation energy condition a is characterized by the monthly decrease in formation pressure: weak (a2>0.5MPa), relatively weak (0.4MPa<a2≤0.5MPa), moderate (0.3MPa<a2≤0.4MPa), relatively strong (0.2MPa<a2≤0.3MPa), and strong (a2≤0.2MPa).

[0046] According to some embodiments of the present invention, the factor levels of the interlayer conditions include: good, relatively good, medium, relatively poor, and poor, which respectively correspond to the levels of good, relatively good, medium, relatively poor, and poor.

[0047] According to the present invention, the interlayer condition b is characterized by the interlayer thickness: good (a3>50m), weak (40<a3≤50m), medium (20<a3≤40m), strong (10<a3≤20m), and strong (a3≤10m).

[0048] According to some embodiments of the present invention, the factor levels of the bottom water communication method include: no connection, partially sealed water; poor communication, partially sealed water; poor communication, connected small water bodies; connected, cracked connection of small water bodies; connected, cracked connection of large water bodies, which respectively correspond to good, relatively good, medium, relatively poor, and poor levels;

[0049] According to some embodiments of the present invention, the factor level of the opening position of the production section includes: upper, middle and bottom, which correspond to good, medium and poor levels, respectively.

[0050] According to some embodiments of the present invention, the factor levels of the opening position of the production section include: upper, upper middle, middle, lower middle and bottom, which correspond to good, better, medium, poor and poor levels, respectively.

[0051] According to some embodiments of the present invention, the method of assigning different weights to each reservoir characteristic index in step b) can be performed using the analytic hierarchy process.

[0052] According to some embodiments of the present invention, the steps of the analytic hierarchy process include: first, comparing the importance of each factor pairwise within each level, and representing the comparison results using an appropriate scale, as shown in Table 2 below, and writing them in matrix form to obtain the judgment matrix U. Then, solving for the largest eigenvalue and the corresponding eigenvector of the judgment matrix to determine the relative importance weights of each indicator at each level.

[0053]

[0054] Table 2. Scale Table for the Analytic Hierarchy Process (AHP)

[0055] 1 Both factors are equally important. 3 Between the two factors, one factor is slightly more important than the other. 5 When comparing two factors, one factor is significantly more important than the other. 7 Between two factors, one factor is more important than the other. 9 Between two factors, one is extremely more important than the other. 2、4、6、8 The median of the two adjacent judgments above reciprocal <![CDATA[The scale for the comparison of factor i and j is ∪ ij , then the scale ∪ for the comparison of factor j and i ij＝ 1 / ∪ ji >

[0056] According to the present invention, steps b)-d) are specifically described as follows:

[0057] For example, if the reservoir characteristic indicators of a certain oil well are determined by a level corresponding to the reservoir type, formation energy conditions, interlayer conditions, bottom water connection method, and production interval opening location, then the reservoir characteristic indicators of the well can be obtained into a 5×5 evaluation matrix according to the evaluation vector in Table 1. The weights of the reservoir type, formation energy conditions, interlayer conditions, bottom water connection method, and production interval opening location constitute a 1×5 weight vector. Multiplying the obtained 1×5 weight vector by the 5×5 evaluation matrix of the well's "reservoir characteristic indicators" yields a 1×5 vector F. The five numbers in the obtained vector comprehensively represent the comprehensive evaluation numerical result of an oil well's "reservoir type, interlayer conditions, formation energy conditions, bottom water connection method, and production interval opening location", and become part of the BP neural network input indicators.

[0058] The second aspect of the present invention also provides the application of the method described in the first aspect of the present invention in the development of carbonate fracture-vuggy bottom water reservoirs, particularly in water cone early warning.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] The present invention provides a method for calculating the critical water cone production rate in carbonate fractured-vuggy bottom-water reservoirs. This method is based on fuzzy clustering analysis and BP neural network methods. By utilizing production and fluid property data from already operational development wells, it combines the advantages of both methods, overcoming the limitations of existing critical water cone production formulas, such as limited applicability and difficulty in parameter selection. This improves the accuracy of critical water cone production rate prediction, fully considers the production dynamics of oil wells at different stages, and more accurately predicts the critical water cone production rate in carbonate fractured-vuggy bottom-water reservoirs. Furthermore, by combining actual production with the critical water cone production rate of oil wells, appropriate water control and oil stabilization measures can be proposed to improve the final recovery rate of carbonate fractured-vuggy bottom-water reservoirs. Attached Figure Description

[0061] Figure 1 This is a flowchart of the method for calculating the critical production rate of the water cone in a fractured-vuggy bottom water reservoir of carbonate rock according to the present invention.

[0062] Figure 2 This is a schematic diagram of the BP artificial neural network process for the critical production rate of water cone in carbonate rock fractured-vuggy bottom water reservoirs according to the present invention. Detailed Implementation

[0063] The present invention will be described in detail below with reference to the embodiments, but the present invention is not limited to the scope disclosed in the following embodiments.

[0064] Example 1

[0065] (I) Twenty-five wells that have been put into production and developed in carbonate fractured-vuggy bottom water reservoirs were selected as sample wells for fuzzy cluster analysis to predict the critical production of the water cone using the fuzzy cluster analysis method. The specific steps are as follows:

[0066] (1) Standardization of raw data from sample wells: The analytical indicators selected for the sample wells are reservoir permeability, crude oil viscosity, crude oil density, crude oil volume factor, oil-water density difference, reservoir thickness, perforation thickness, average production, and oil pressure. The original matrix is ​​obtained, and the raw data of the selected analytical indicators are normalized to the closed interval [0, 1]. The raw data of each analytical indicator are standardized, and the specific method is as follows:

[0067] a. For the "proliferation" indicator, the data standardization method of Formula I is adopted as follows:

[0068]

[0069] b. For the "minimizing" indicator, the data standardization method of Equation II is adopted as follows:

[0070]

[0071] Among them, X ij X′ represents the standardized value of the original data of well j in sample i. ij X represents the actual value of the original data of the index of sample well j in sample i. jmin Represents the minimum value of the j-th index parameter, X jmax This represents the maximum value among the j-th index parameters;

[0072] (2) Constructing a similarity matrix: Based on the obtained standardized data, the similarity coefficients between various indicators of each sample well are calculated using the cosine of the included angle method, and a fuzzy similarity matrix is ​​constructed.

[0073] (3) Fuzzy clustering of sample wells: Based on the fuzzy similarity matrix in step (2), the sample wells are classified using MATLAB software. The threshold λ is set to 0.9, so that the 25 sample wells are divided into classes I, II, III, IV and V, as shown in Table 3 below.

[0074] Table 3 Cluster analysis of sample well classification

[0075]

[0076]

[0077] (4) Then, calculate the similarity coefficient between each predicted well and the sample wells, identify the sample well with the highest similarity to the predicted well, and classify the predicted well into that sample well category. After determining the sample well category, use the average critical water cone production of that category of sample wells as the critical production (Q) of the predicted well. 模糊聚类The results are shown in Table 4 below. By continuously increasing the number of sample wells, enriching the sample set, and increasing the value of λ, the sample set can be divided into more classes, thereby increasing the accuracy of fuzzy clustering analysis in calculating critical yield.

[0078] Table 4. Critical yield prediction results using fuzzy clustering analysis.

[0079]

[0080] (II) Select the above sample wells for BP artificial neural network critical production prediction (see details for specific operation). Figure 2 )

[0081] Using the parameters of the sample wells for critical production prediction using the BP artificial neural network in Table 5, the accuracy of the model was improved by adjusting the number of hidden layers and nodes in the BP artificial neural network. Ultimately, the artificial neural network was determined to contain two hidden layers: the first layer with 45 nodes and the second layer with 10 nodes, achieving a fitting accuracy of 99%.

[0082] The established BP artificial neural network model was used to calculate the critical water cone production (Q) of 15 prediction wells, thereby obtaining the critical water cone production of the BP artificial neural network. 神经网络 The prediction results are shown in Table 6.

[0083] (III) For 15 predictive wells, the critical water cone production rate (0.43 × Q) was calculated using the method of this invention. 模糊聚类 +0.57×Q 神经网络 The calculation results were compared with those of the Dupuit method, Chaperon method, Schols method, Meyer-Garder method, Hoyland method, S&C method, B&J method, Renqiu water cone formula method, and diaphragm method. The results are shown in Table 7.

[0084] As shown in Table 7, the critical production rate of the water cone calculated by the formula method deviates significantly from the stable production rate of the oil well. This is mainly due to the limited applicability of each formula and the difficulty in determining the values ​​of the key parameters. The water cone critical production rate calculation method of this invention effectively avoids these problems, and the calculated critical production rate of the water cone has a high consistency rate with the stable production rate of the oil well.

[0085]

[0086]

[0087]

[0088]

Claims

1. A method for calculating the critical production rate of the water cone in a carbonate fractured-vuggy bottom water reservoir, the calculation method comprising: The critical water cone production Q of carbonate fractured-vuggy bottom water reservoirs was calculated by combining fuzzy clustering analysis and BP artificial neural network method. 临界产量 , where Q 临界产量 = a×Q 模糊聚类 + (1-a)×Q 神经网络 , a∈[0,1], Q 模糊聚类 Q is the critical yield obtained based on fuzzy clustering analysis. 神经网络 The critical yield obtained using the BP artificial neural network method is determined by performing regression analysis and practical verification on the critical yield values ​​of sample wells with known critical yields and the critical yield values ​​of the sample wells calculated using fuzzy clustering analysis and the BP artificial neural network method, respectively. The fuzzy clustering analysis method includes the following steps: (1) Standardization of raw data of sample wells: wells that have been put into production and development in carbonate fractured-vuggy bottom water reservoirs are selected as sample wells, and the analysis indicators of the sample wells are selected to obtain the original matrix. The raw data of the analysis indicators are normalized to the closed interval [0, 1] to obtain the standardized data of the raw data of each analysis indicator of the sample wells. (2) Constructing a similarity matrix: Based on the obtained standardized data, calculate the similarity coefficients between each index of each sample well and construct a fuzzy similarity matrix; (3) Fuzzy clustering of sample wells: Based on the fuzzy similarity matrix in step (2), the sample wells are classified using MATLAB software; (4) Classification of target wells and prediction of critical production: Based on the data of the target wells corresponding to the analysis indicators, the similarity coefficient between the target wells and the sample wells is calculated to obtain the sample wells with the highest similarity to the target wells. At the same time, the target wells are classified into the sample well category. The average critical production of the water cone of the sample wells in this category is the critical production of the water cone of the target well, Q. 模糊聚类 ; The BP artificial neural network method includes the following steps: 1) Determine the input and output indicators of the BP artificial neural network: the input indicators include the reservoir characteristic indicators of the sample wells and the production dynamic indicators of the sample wells, and the output indicator is the critical production rate of the oil well; 2) Input index digitization: The input indexes of the sample wells are digitized to obtain the input vector of the sample wells; 3) Optimization of BP artificial neural network model structure: By adjusting the number of hidden layers and the number of nodes in each layer, the structure of the BP artificial neural network model is optimized by using trial and error comparison to determine the number of hidden layers and the number of nodes in each layer; 4) Training the BP artificial neural network model: Divide the sample well into n parts, one part is the test sample and the remaining n-1 parts are the training samples. Normalize the input and output data, give the weights and thresholds of each layer, train the samples and adjust the corresponding parameters until the model meets the error requirements, and obtain the BP artificial neural network model. 5) Target well critical water cone production prediction: Based on the trained BP artificial neural network model, the corresponding input index of the target well is input to obtain the predicted critical water cone production Q of the target well. 神经网络 .

2. The calculation method according to claim 1, characterized in that, Q 临界产量 = 0.43×Q 模糊聚类 + 0.57×Q 神经网络 。 3. The calculation method according to claim 1 or 2, characterized in that, The original matrix is: , Where N represents the number of sample wells, and K represents the number of analytical indicators for each sample well.

4. The calculation method according to claim 3, characterized in that, N is not less than 10.

5. The calculation method according to claim 1 or 2, characterized in that, The analytical indicators include: reservoir permeability, crude oil viscosity, crude oil density, crude oil volume factor, oil-water density difference, reservoir thickness, perforation thickness, average production, and oil pressure.

6. The calculation method according to claim 1 or 2, characterized in that, In step (2), the similarity coefficient is calculated using the cosine of the included angle method. , where r ij x represents the similarity coefficient between sample well i and oil well j; ik ' represents the normalized value of index k for sample well i; x jk ' represents the normalized value of index k for sample well j.

7. The calculation method according to claim 1 or 2, characterized in that, In step (3), the maximum tree method is used to classify the sample wells by inputting the horizontal threshold λ∈[0,1].

8. The calculation method according to claim 1 or 2, characterized in that, In step 1) of the BP artificial neural network method, the reservoir characteristic indicators of the sample well include reservoir type, interlayer conditions, formation energy conditions, opening location of production intervals, and bottom water connection method. And / or the production dynamic indicators of the sample wells include production dynamic indicators during the dry period, low water cut period, medium water cut period, and high water cut period.

9. The calculation method according to claim 8, characterized in that, The production dynamic indicators of the sample wells include production time during the dry period, average production during the dry period, production time during the low water cut period, average production during the low water cut period, and average water cut during the low water cut period; production time during the medium water cut period, average production during the medium water cut period, and average water cut during the medium water cut period; and production time during the high water cut period, average production during the high water cut period, and average water cut during the high water cut period.

10. The calculation method according to claim 8, characterized in that, In step 2) of the BP artificial neural network method, the reservoir characteristic indicators are digitized, which is done through a method including the following steps: a) The different effects of the reservoir characteristic indicators on the critical production of the water cone are divided into five evaluation levels: good, relatively good, medium, poor, and poor. Each evaluation level corresponds to an evaluation vector consisting of 5 numbers, the sum of which is 1. The evaluation vector is determined by combining the experience of those skilled in the art and expert scoring. b) Assign different weights to each reservoir characteristic index to obtain the reservoir characteristic index weight vector. c) Determine the evaluation level corresponding to the reservoir characteristic index of the oil well, and thus obtain the evaluation vector matrix corresponding to the oil well; d) Multiply the evaluation vector matrix by the weight vector to obtain the digital results of each reservoir characteristic index of the target well. In step c), the evaluation level of the target oil well is determined by dividing each reservoir characteristic index into five factor levels according to its different impacts on the critical production rate of the water cone. These five factor levels correspond to the grades of good, relatively good, moderate, relatively poor, and poor, respectively. The factor levels of the reservoir types include: cave type, small cave type, fissure-cavity type, small fissure type, and fissure type, which correspond to good, relatively good, medium, poor, and very poor levels, respectively; The factor levels of the formation energy conditions include: weak, relatively weak, moderate, relatively strong, and strong, which correspond to good, relatively good, moderate, relatively poor, and poor levels, respectively. The factor levels of the isolation conditions include: good, relatively good, medium, relatively poor, and poor, which correspond to the levels of good, relatively good, medium, relatively poor, and poor, respectively. The factor levels of the bottom water communication methods include: no connection, partially sealed water; poor communication, partially sealed water; poor communication, connected small water bodies; connected, cracked connection of small water bodies; connected, cracked connection of large water bodies, which correspond to good, relatively good, medium, relatively poor, and poor levels, respectively. The factor levels for the opening position of the production section include: upper, upper middle, middle, lower middle and bottom, which correspond to good, better, medium, poor and poor levels, respectively.

11. The calculation method according to claim 10, characterized in that, The evaluation vector is: {Good: 0.83, 0.16, 0.01, 0.00, 0.00;} Good: 0.16, 0.68, 0.14, 0.02, 0.00; Medium: 0.02, 0.27, 0.42, 0.27, 0.02; Poor: 0.00, 0.02, 0.14, 0.68, 0.16; Difference: 0.00, 0.00, 0.01, 0.16, 0.83}.

12. The application of the method as described in any one of claims 1-11 in the development of carbonate fracture-vuggy bottom water reservoirs.

13. The application according to claim 12, characterized in that, The application of the method in water cone early warning.

Images