A method for determining the main controlling factors of flow field regulation in old water flooding reservoirs based on data-driven

Abstract

The present invention provides a method for determining the main controlling factors of flow field regulation in old water-driven oil reservoirs based on data-driven, and the method comprises: step 1, acquiring data and performing quality inspection; step 2, associating and fusing oil and water well data; step 3, establishing characteristic parameters for flow field description; step 4, establishing flow field adjustment effective labels; step 5, establishing main controlling factor analysis sample data; step 6, determining the main controlling factors based on association rule analysis. The method for determining the main controlling factors of flow field regulation in old water-driven oil reservoirs based on data-driven determines the main influencing factors of flow field regulation in water-driven oil reservoirs in the ultra-high water cut period and gives the effective regulation direction, providing technical support for further improving the recovery rate of water-driven oil reservoirs in the ultra-high water cut period.

Description

Technical Field

[0001] The invention relates to the technical field of oilfield development, and in particular to a method for determining main control factors of flow field regulation of old water-driven oil reservoirs based on data. Background Art

[0002] Flow field control is the main means to further improve the recovery rate of old oil fields. However, due to the complexity of flow field control factors and driving mechanisms, the control effects of different reservoirs vary significantly. How to find the main control factors, clarify the direction of flow field control, and achieve accurate and effective flow field control is an urgent problem to be solved. However, due to the complex underground seepage laws of old oil fields, the flow field control factors include both static factors and dynamic factors, with many types, complex relationships, and large amounts of data. Manual analysis is extremely difficult, making it difficult to find the main control factors using conventional analysis methods. Big data intelligent analysis technology has the ability to mine "hidden" relationships and main control factors, providing a new way to study the main influencing factors of reservoir flow field control.

[0003] In the Chinese patent application with application number: CN201910431091.7, a method for regulating the flow field of an oil reservoir is involved, and the method comprises: obtaining the oil production ratio between each water injection well and the adjacent oil production well in the target oil reservoir, and the oil production ratio is: the mass ratio of the oil production contribution of the water injection well to the oil production well and the oil production of the oil production well in unit time; obtaining the water injection ratio between each water injection well and the adjacent oil production well in the target oil reservoir, and the water injection ratio is: the mass ratio of the water delivery of the water injection well to the oil production well and the total water injection of the water injection well in unit time; based on the oil production ratio between each water injection well and the adjacent oil production well, and the water injection ratio between each water injection well and the adjacent oil production well, calculating the contribution rate of each water injection well to the adjacent oil production well; and regulating the flow field of the target oil reservoir based on the contribution rate of each water injection well to the adjacent oil production well.

[0004] In the Chinese patent application with application number: CN201710803024.4, a method for designing plane flow field control of water-drive reservoir is involved, including: step 1, analyzing the oil-water boundary flow field, and calculating the pressure field and velocity field distribution of single-phase flow under the conditions of fixed well network and stable injection and production volume; step 2, discretizing the oil-water boundary and the near-well flow boundary of the production well, and calculating the ideal velocity vector and the actual velocity vector of the discrete position; step 3, obtaining the evaluation coefficient, and calculating the difference between the ideal velocity and the actual velocity of the boundary flow field and the single well flow respectively; step 4, judging whether the adjusted evaluation coefficient meets the expected value. When the evaluation coefficient meets the expected value, the process enters step 5; step 5, obtaining the optimal well network and injection and production control design scheme according to the result of the evaluation coefficient.

[0005] In the Chinese patent application with application number: CN202010262704.1, a method for evaluating the adaptability of the flow field in a high water-cut reservoir is involved. The method for evaluating the adaptability of the flow field in a high water-cut reservoir includes: step 1, establishing a numerical simulation model of the reservoir and obtaining evaluation parameters; step 2, performing data preprocessing; step 3, calculating the evaluation index of the adaptability of the flow field in a high water-cut reservoir; step 4, finding the optimal control plan for the reservoir.

[0006] The above existing technologies are greatly different from the present invention and fail to solve the technical problem we want to solve. For this reason, we have invented a new data-driven method for determining the main controlling factors of flow field regulation in old water flooding reservoirs. Summary of the invention

[0007] The purpose of the present invention is to provide a data-driven method for determining the main controlling factors of flow field regulation in water-flooded old oil reservoirs, which provides technical support for further improving the recovery rate of water-flooded oil reservoirs in the ultra-high water-cut period.

[0008] The purpose of the present invention can be achieved through the following technical measures: a data-driven method for determining the main controlling factors of flow field regulation in old water-flooded oil reservoirs, the method comprising: step 1, acquiring data and performing quality inspection; step 2, associating and fusing oil and water well data; step 3, establishing flow field description characteristic parameters; step 4, establishing flow field adjustment effectiveness labels; step 5, establishing main controlling factor analysis sample data; step 6, determining the main controlling factors based on association rule analysis.

[0009] The purpose of the present invention can also be achieved by the following technical measures:

[0010] In step 1, for the target study area, the following raw data are obtained from the database: well location parameters, geological sub-layer comparison data, layer physical property parameters, oil and water well perforation history data, monthly production dynamic data, and monthly water injection dynamic data.

[0011] In step 1, after data acquisition, histograms and box plots are used to perform outlier analysis on the original data, and outliers are eliminated based on the analysis results.

[0012] In step 2, firstly, the water wells in the same layer as the oil well during the production period of the oil well are determined based on the perforation data and the small layer comparison data, and the water injection volume of the water wells is split. Then, the associated water wells of the oil well are determined through the correlation analysis of the oil and water wells, and the oil and water well data are fused.

[0013] In step 2, the fusion of oil and water well data is to add the injection volume data of the determined associated water wells along the time axis and put it into the sample data of the oil well as an attribute.

[0014] In step 2, the correlation analysis method of oil and water wells refers to the selection of a multivariate linear regression model (1), with weights β kj represents the degree of connectivity between the production well and the injection well k, β 0j represents the injection-production imbalance coefficient of water-drive reservoir;

[0015] The liquid production of the jth production well is q j It is represented by the following multiple linear regression model:

[0016]

[0017] in, represents the liquid production of the jth oil well at time step t; i k (t) is the injection volume of the kth injection well at time step t; β kj Represents the multiple linear regression weights of the j-th oil well and the k-th injection well.

[0018] In step 4, the weight factor β of the multivariate linear regression model is kj The minimum residual square sum SSE of the observed and estimated values ​​of the production volume of the production well is obtained by using the least squares method.

[0019]

[0020] Where T is the number of data in the calculation time domain; the least squares estimate of the minimum value obtained by SSE should satisfy the following equation:

[0021]

[0022] This equation can also be expressed as:

[0023]

[0024] In the formula, Cov ih is the covariance between the injection volume of injection well i and injection well h; Cov ij The covariance between the injection rate of injection well i and the production rate of production well j;

[0025] Then equation (4) can be expressed by the following linear equations:

[0026]

[0027] In equation (5), the first part of the matrix is ​​the covariance matrix of the injection data vector, which is obtained through the injection dynamic data; the column vector of the second part is the least squares estimate of the weight of the multivariate linear regression model of the injection-production system; the right side of the equation is the covariance column vector between the injection volume of the injection well and the liquid production volume of the production well, which is obtained through the injection-production dynamic data;

[0028] The correlation between the estimated value of liquid production after regression and the observed value is expressed by the coefficient of determination R 2 To express:

[0029]

[0030]

[0031] Where SSE is the residual sum of squares, and SSY is the total sum of squares.

[0032] In step 3, based on the collected oil and water well field data, a flow field description feature parameter set is established in combination with the reservoir engineering analysis method and the flow field description requirements, including: the well number of the oil well, production time, bottom hole flow pressure, daily liquid production, daily oil production, daily water production, oil production index, liquid production index, specific oil production index, specific liquid production index, water cut rise rate, monthly liquid production change, monthly oil production change, effective thickness change, production layer change, monthly production days change, cumulative production time, daily water injection volume of the water well, water absorption index, specific water absorption index, water absorption index change, monthly water injection change, effective thickness change, water injection layer change, water injection days change, and cumulative water injection time.

[0033] In step 4, the goal of flow field adjustment is to control water and stabilize oil production or to reduce water and increase oil production. The evaluation parameters for flow field adjustment are determined to be two indicator parameters: monthly water cut increase rate of a single well and daily oil-water change of a single well. The data distribution characteristics of the evaluation indicator parameters of the target reservoir are statistically analyzed to determine the evaluation standards, and the samples are labeled according to the standards.

[0034] In step 5, the characteristic parameters describing the flow field have multi-dimensional and multi-valued characteristics, and the data are complex and diverse. The continuous data of different attributes are normalized or discretized accordingly in combination with the physical meaning of the data to form sample data.

[0035] In step 6, the association analysis method based on the Apriori algorithm is used to mine the association rules that lead to the effectiveness of flow field regulation, and the main control parameters are determined according to the support and confidence of the rules.

[0036] The data-driven method for judging the main controlling factors of flow field regulation in old water-flooded oil reservoirs in the present invention includes a method for associating and fusion of multidimensional data of oil and water wells based on multiple linear regression, establishment of characteristic parameters for describing the effectiveness of flow field regulation, establishment of effective labels, and judgment of main controlling factors based on association analysis. This method uses a method for associating and fusion of multidimensional data of oil and water wells based on multiple linear regression to achieve the associative fusion of oil and water well data. On this basis, a basic learning sample for judging the main controlling factors of water well data is formed by establishing characteristic parameters for describing the effectiveness of flow field regulation and establishing effective labels. Finally, the main controlling factor judgment method based on association analysis of the present invention is used to determine the main influencing factors of flow field regulation in water-flooded oil reservoirs in the ultra-high water-cut period and give the direction of effective regulation, so as to provide technical support for further improving the recovery rate of water-flooded oil reservoirs in the ultra-high water-cut period. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flow chart of a specific embodiment of the method for determining the main controlling factors of flow field regulation in old water flooding reservoirs based on data-driven according to the present invention;

[0038] Figure 2 It is a schematic diagram of a common layer water well in a certain oil well W-PRO production cycle in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0039] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0040] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations and / or combinations thereof.

[0041] like Figure 1 As shown, Figure 1 The flowchart of the method for determining the main controlling factors of flow field regulation of water flooding old oil reservoirs based on data drive of the present invention. The method for determining the main controlling factors of flow field regulation of water flooding oil reservoirs in the ultra-high water cut period driven by multi-dimensional data includes the following steps:

[0042] Step 1: Get data and perform quality check;

[0043] For the target study area, the following raw data were obtained from the database: well location parameters, geological sub-layer comparison data, layer physical property parameters, oil and water well perforation history data, monthly production dynamic data, and monthly water injection dynamic data. After the data were obtained, histograms and box plots were used to analyze the raw data for outliers, and outliers were eliminated based on the analysis results.

[0044] Step 2: oil and water well data association and fusion;

[0045] Firstly, the water wells in the same layer as the oil wells during the production period of the oil wells are determined based on the perforation data and small layer comparison data, and the water injection volume of the water wells is split. Then, the associated water wells of the oil wells are determined through the correlation analysis of the oil and water wells, and the oil and water well data are fused.

[0046] The correlation analysis method of oil and water wells refers to the use of multiple linear regression model (1), weight β kj represents the degree of connectivity between the production well and the injection well k, β 0j Represents the injection-production imbalance coefficient of water-drive reservoir.

[0047] The liquid production of the jth production well is q j It can be represented by the following multiple linear regression model:

[0048]

[0049] in, represents the liquid production of the jth oil well at time step t; i k (t) is the injection volume of the kth injection well at time step t; β kj Represents the multiple linear regression weights of the j-th oil well and the k-th injection well.

[0050] Weight factor β of the multivariate linear regression model kj The least square method is used to find the minimum residual sum of squares (SSE) between the observed and estimated values ​​of the production volume of the production well.

[0051]

[0052] Where T is the number of data in the calculation time domain. The least squares estimate of SSE to obtain the minimum value should satisfy the following equation:

[0053]

[0054] This equation can also be expressed as:

[0055]

[0056] In the formula, Cov ih is the covariance between the injection volume of injection well i and injection well h; Cov ijThe covariance between the injection rate of injection well i and the production rate of production well j.

[0057] Then equation (4) can be expressed by the following linear equations:

[0058]

[0059] In equation (5), the first part of the matrix is ​​the covariance matrix of the injection data vector, which can be obtained through the injection dynamic data; the column vector of the second part is the least squares estimate of the weight of the multivariate linear regression model of the injection-production system; the right side of the equation is the covariance column vector between the injection volume of the injection well and the liquid production of the production well, which can be obtained through the injection-production dynamic data.

[0060] The correlation between the estimated value of liquid production after regression and the observed value is expressed as the coefficient of determination R 2 To express:

[0061]

[0062]

[0063] Where SSE is the residual sum of squares, and SSY is the total sum of squares.

[0064] The fusion of oil and water well data is to add the injection volume data of the determined related water wells along the time axis and put it into the sample data of the oil well as an attribute.

[0065] Step 3, establishing characteristic parameters for flow field description;

[0066] Based on the collection of oil and water well field data, a flow field description feature parameter set is established in combination with reservoir engineering analysis methods and flow field description requirements, including: well number, production time, bottom hole flowing pressure, daily liquid production, daily oil production, daily water production, oil production index, liquid production index, specific oil production index, specific liquid production index, water cut rise rate, monthly liquid production change, monthly oil production change, effective thickness change, production layer change, monthly production days change, cumulative production time; daily water injection volume, water absorption index, specific water absorption index, water absorption index change, monthly water injection change, effective thickness change, water injection layer change, water injection days change, and cumulative water injection time of water wells.

[0067] Step 4, establishing flow field adjustment effect label;

[0068] Taking "controlling water to stabilize oil production or reducing water to increase oil production" as the goal of flow field adjustment, the evaluation parameters for flow field adjustment are determined to be two index parameters: monthly water cut increase rate of a single well and daily oil-water change of a single well. The distribution characteristics of the evaluation index parameter data of the target reservoir are statistically analyzed, the evaluation standard is determined, and the samples are labeled according to the standard.

[0069] Step 5, establish sample data for main control factor analysis;

[0070] The characteristic parameters describing the flow field have multi-dimensional and multi-value characteristics, and the data is complex and diverse. It is necessary to normalize or discretize the continuous data of different attributes in accordance with the physical meaning of the data and the data format requirements of the machine learning algorithm. In this patent, continuous variables need to be discretized. When discretizing, the appropriate binning method should be selected according to the physical meaning of the data. Finally, all the sorted feature data and label data are integrated to form sample data.

[0071] Step 6: Determine the main controlling factors based on association rule analysis.

[0072] The association analysis method based on Apriori algorithm is used to mine the association rules that lead to the effectiveness of flow field regulation, and the main control parameters are determined according to the support and confidence of the rules.

[0073] In the specific embodiment 1 of the present invention, an application study is carried out on a water injection development reservoir in Shengli Oilfield. The development has been going on for 35 years. In the past, there were 20 oil wells, 15 water wells, and 5 small layers put into production. The main controlling factors of flow field adjustment are studied by the method of the present invention: the process is as follows: Figure 1 express:

[0074] Step 1, data acquisition and quality check, 4483 original monthly production data of the study area were collected, and after outlier and missing value monitoring, 904 outlier and missing value records were removed, leaving 3579 complete record data.

[0075] Step 2: Fusion of oil and water well data: Determine the water wells in the same layer as the oil wells during the production period of the oil wells based on the perforation data and the small layer comparison data (see Figure 2 ), and split the water injection volume of the water well, and then determine the associated water wells of the oil well through the correlation analysis of the oil and water wells. For example, according to the correlation analysis result Table 1, the associated water wells of the oil well W-PRO are W-INJ10, W-INJ1, W-INJ3 and W-INJ6. Then, the data of the four associated water wells are weighted summed and fused with the characteristic parameters of the oil well.

[0076] Table 1 Schematic diagram of the results of the oil-water well correlation analysis

[0077]

[0078]

[0079] Step 3, establish a flow field description characteristic parameter set, including: oil well number, production time, bottom hole pressure, daily liquid production, daily oil production, daily water production, oil production index, liquid production index, specific oil production index, specific liquid production index, water cut rise rate, monthly liquid production change, monthly oil production change, effective thickness change, production layer change, monthly production days change, cumulative production time, daily water injection volume of water wells, water absorption index, specific water absorption index, water absorption index change, monthly water injection change, effective thickness change, water injection layer change, water injection days change, and cumulative water injection time.

[0080] Step 4: Statistically analyze the data distribution characteristics of the two indicator parameters, namely, the monthly water cut increase rate of the target reservoir and the daily oil-water change of the target reservoir, determine the evaluation criteria (see Table 2), and establish label data.

[0081] Table 2 Parameters and standards for judging the effectiveness of flow field control in a certain oil reservoir

[0082]

[0083] Step 5: After normalizing or discretizing the continuous attribute data accordingly (see Table 3 for examples of continuous attribute discretization), the organized data are integrated to form sample data.

[0084] Table 3 Example of continuous attribute discretization

[0085]

[0086] Step 6, using the association analysis method based on the Apriori algorithm to mine the main association rules for oil well dewatering and oil production increase, the main influencing parameters of oil well dewatering and oil production increase are determined according to the support and confidence of the rules, such as the cumulative water injection volume of the associated water wells, liquid extraction measures, monthly water injection volume changes, effective thickness changes, cumulative production time, and specific liquid production index.

[0087] In the specific embodiment 2 of the present invention, an application study was conducted on a water-flooding reservoir in Shengli Oilfield. The reservoir was put into production in February 1973, and natural energy was developed from February 1973 to February 1976. Water flooding was carried out from March 1976 to February 1997, and polymer injection was carried out from March 1997 to April 2002. In April 2002, the entire area was transferred to subsequent water flooding. The period before polymer injection development was selected for analysis. The development lasted for 24 years, with a comprehensive water content of 95%, 60 oil wells, 42 water wells, and 9 small layers put into production. The main controlling factors for adjusting the water-flooding flow field at different water-content stages were studied by the method of the present invention. The process is as follows: Figure 1 express:

[0088] Step 1: Data acquisition and quality inspection in the water injection development stage. 18,463 pieces of original monthly production data of the study area were collected, and 4 characteristic parameters with a large number of missing or invalid data were removed. After monitoring of outliers and missing values, 3,372 records of outliers and missing values ​​remained, and 15,091 complete records remained.

[0089] Step 2, fusion of oil and water well data, determine the water wells in the same layer as the oil well during the production period of the oil well based on the perforation data and the small layer comparison data, and split the water injection volume of the water wells, and then determine the associated water wells of the oil wells through correlation analysis of the oil and water wells. For example, according to the correlation analysis results in Tables 4-5, the associated water wells of the oil well GDX-POR1 are GDX-W1, GDX-W2, and GDX-W5, and the associated water wells of the oil well GDX-POR3 are GDX-W1, GDX-W2, GDX-W3, and GDX-W5. Then, the data of the associated water wells of each oil well are weighted summed and fused with the corresponding oil well characteristic parameters.

[0090] Table 4 Correlation analysis results between GDX-POR1 in oil and water wells

[0091] GDX-POR1 Correlation coefficient Standard error Significance test GDX-W1 0.194 -0.213 0.381 GDX-W2 0.235 0.142 0.272 GDX-W5 0.237 -0.37 0.293

[0092] Table 5 Correlation analysis results between GDX-POR2 of oil and water wells

[0093] GDX-POR2 Correlation coefficient Standard error Significance test GDX-W1 0.022 -0.365 0.092 GDX-W2 0.036 0.121 0.543 GDX-W3 0.057 -0.406 0.113 GDX-W5 0.042 -0.157 0.535

[0094] Step 3, based on the valid data, establish a flow field description characteristic parameter set, including: oil well number, production time, daily liquid production, daily oil production, daily water production, water cut rise rate, monthly liquid production change, monthly oil production change, monthly production days change, cumulative production time, daily water injection volume of water wells, water absorption index, monthly water injection change, effective thickness change, water injection layer change, water injection days change, cumulative water injection time.

[0095] Step 4: Statistically analyze the data distribution characteristics of the two indicator parameters, namely, the monthly water cut increase rate of the target reservoir and the daily oil-water change of the target reservoir, determine the evaluation criteria (see Table 6), and establish label data.

[0096] Table 6 Parameters and standards for judging the effectiveness of flow field control in a certain oil reservoir

[0097]

[0098] Step 5: After normalizing or discretizing the continuous attribute data (see Table 3 for examples of continuous attribute discretization), since this example wants to analyze whether the main controlling factors of different water content stages are different, an attribute parameter of the water content stage is added (see Table 7 for definition), and the integrated and sorted data are formed into sample data.

[0099] Table 7 Example of continuous attribute discretization

[0100]

[0101] Step 6, using the association analysis method based on the Apriori algorithm, it is found that the regulation effect of the low water-cut period block is mainly reflected in the aspects of water reduction and oil production increase and water control and oil stabilization, and the main influencing parameters of oil well water reduction and oil production increase are layer changes, effective thickness changes, production time changes, number of oil and water wells, and liquid production index parameters. The regulation effect of the ultra-high water-cut stage is mainly reflected in water reduction and oil production increase and water control and oil production increase. According to the support and confidence of the association rules, it is inferred that the main influencing factors are liquid volume changes (liquid lifting measures), monthly water injection volume changes, effective thickness changes, cumulative production time, and specific liquid production index. It shows that the parameters that are effective in the low water-cut period represent the plugging and filling measures of oil and water wells and the new well investment measures, and the parameters that are effective in the ultra-high water-cut period represent the control methods such as liquid lifting, plugging and filling holes, and water injection allocation, which is consistent with the reservoir understanding.

[0102] In the specific embodiment 3 of the present invention, the application research is carried out for the subsequent water flooding of a water flooding development oil reservoir in Shengli Oilfield. After 16 years of development, the comprehensive water content has increased from 91.3% to 98.2%. It is currently in the late stage of ultra-high water content. There are 59 oil wells and 42 water wells, which are divided into two sets of upper and lower layers for exploitation. The main controlling factors for adjusting the water flooding flow field at different water content stages are studied by the method of the present invention. The process is as follows: Figure 1 express:

[0103] Step 1: Data acquisition and quality check. There are 16,346 original monthly production data in the study area. Four characteristic parameters with a large number of missing or invalid data are removed. After outlier and missing value monitoring, there are 4,830 records of outliers and missing values ​​remaining, and 11,516 complete records remaining.

[0104] Step 2, fusion of oil and water well data, determine the water wells in the same layer as the oil well during the production period of the oil well based on the perforation data and small layer comparison data, and split the water injection volume of the water wells, and then determine the associated water wells of the oil well through correlation analysis of the oil and water wells, and finally perform weighted summation of the data of the associated water wells and fuse them with the characteristic parameters of the oil well.

[0105] Step 3, based on the valid data, establish a flow field description characteristic parameter set, including: oil well number, production time, daily liquid production, daily oil production, daily water production, water cut rise rate, monthly liquid production change, monthly oil production change, monthly production days change, cumulative production time, daily water injection volume of water wells, water absorption index, monthly water injection change, effective thickness change, water injection layer change, water injection days change, cumulative water injection time.

[0106] Step 4: Statistically analyze the data distribution characteristics of the two indicator parameters, namely, the monthly water cut increase rate of the target reservoir and the daily oil-water change of the target reservoir, determine the evaluation criteria (see Table 8), and establish label data.

[0107] Table 8 Parameters and standards for judging the effectiveness of flow field control in a certain oil reservoir

[0108]

[0109] Step 5: After normalizing or discretizing the continuous attribute data, the organized data is integrated to form sample data.

[0110] Step 6: The water content increase rate and oil production increase of a single well in the late stage of ultra-high water content reservoir are relatively small. The samples of "oil production increase" and "water control" are taken as effective samples. The main association rules are mined by association analysis method. According to the support and confidence of the rules, the main influencing parameters are determined to be the cumulative water injection volume of the associated water wells, the change of the associated water wells, the change of water injection volume, the liquid lifting measures, the change of effective thickness, and the cumulative production time.

[0111] The data-driven method for determining the main controlling factors of flow field regulation in old water-flooded oil reservoirs in the present invention gives full play to the multi-dimensional and multi-scale data advantages of old oil fields, carries out research on the identification of main controlling factors of flow fields based on machine learning, and helps the continuous and efficient development of old oil fields.

[0112] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions recorded in the aforementioned embodiments or replace some of the technical features therein with equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0113] Except for the technical features described in the specification, all other technical features are known technologies to those skilled in the art.

Claims

1. A data-driven method for determining the main controlling factors of flow field regulation in old water-flooded reservoirs, characterized in that: The data-driven method for determining the main controlling factors of flow field regulation in old water flooding reservoirs includes: Step 1: Get data and perform quality check; Step 2, associating and fusing the oil and water well data; Step 3, establishing characteristic parameters for describing the flow field; Step 4, establishing a flow field adjustment effective label; Step 5, establish sample data for main control factor analysis; Step 6, determine the main controlling factors based on association rule analysis; In step 2, firstly, the water wells in the same layer as the oil wells during the production period of the oil wells are determined based on the perforation data and the small layer comparison data, and the water injection volume of the water wells is split. Then, the associated water wells of the oil wells are determined through the correlation analysis of the oil and water wells, and the oil and water wells data are fused; The fusion of oil and water well data is to add the injection volume data of the determined related water wells along the time axis and put it into the sample data of the oil well as an attribute; The correlation analysis method of oil and water wells refers to the use of multiple linear regression model, weight β kj represents the degree of connectivity between the production well and the injection well k, β 0j represents the injection-production imbalance coefficient of water-drive reservoir; The liquid production of the jth production well is q j It is represented by the following multiple linear regression model: in, represents the liquid production of the jth oil well at time step t; i k (t) is the injection volume of the kth injection well at time step t; β kj represents the multivariate linear regression weights of the j-th oil well and the k-th injection well; Weight factor β of the multivariate linear regression model kj The minimum residual square sum SSE of the observed and estimated values ​​of the production volume of the production well is obtained by using the least squares method. Where T is the number of data in the calculation time domain; the least squares estimate of the minimum value obtained by SSE should satisfy the following equation: The equation is expressed as follows: In the formula, Cov ih is the covariance between the injection volume of injection well i and injection well h; Cov ij The covariance between the injection rate of injection well i and the production rate of production well j; Then equation (4) is expressed by the following linear equations: In equation (5), the first part of the matrix is ​​the covariance matrix of the injection data vector, which is obtained through the injection dynamic data; the column vector of the second part is the least squares estimate of the weight of the multivariate linear regression model of the injection-production system; the right side of the equation is the covariance column vector between the injection volume of the injection well and the liquid production volume of the production well, which is obtained through the injection-production dynamic data; The correlation between the estimated value of liquid production after regression and the observed value is expressed by the coefficient of determination R 2 To express: Where SSE is the residual sum of squares, and SSY is the total sum of squares.

2. The method for determining the main controlling factors of flow field regulation in old water flooding reservoirs based on data-driven according to claim 1 is characterized in that: In step 1, for the target study area, the following raw data are obtained from the database: well location parameters, geological sub-layer comparison data, layer physical property parameters, oil and water well perforation history data, monthly production dynamic data, and monthly water injection dynamic data.

3. The method for determining the main controlling factors of flow field regulation in old water flooding reservoirs based on data-driven according to claim 2 is characterized in that: In step 1, after data acquisition, the original data is analyzed for outliers using histograms and box plots, and outliers are removed based on the analysis results.

4. The method for determining the main controlling factors of flow field regulation in old water flooding reservoirs based on data-driven according to claim 1 is characterized in that: In step 3, based on the collected oil and water well field data, a flow field description feature parameter set is established in combination with the reservoir engineering analysis method and the flow field description requirements, including: the well number of the oil well, production time, bottom hole pressure, daily liquid production, daily oil production, daily water production, oil production index, liquid production index, specific oil production index, specific liquid production index, water cut rise rate, monthly liquid production change, monthly oil production change, effective thickness change, production layer change, monthly production days change, cumulative production time, daily water injection volume of the water well, water absorption index, specific water absorption index, water absorption index change, monthly water injection change, effective thickness change, water injection layer change, water injection days change, and cumulative water injection time.

5. The method for determining the main controlling factors of flow field regulation in old water flooding reservoirs based on data-driven according to claim 1 is characterized in that: In step 4, the goal of flow field adjustment is to control water and stabilize oil production or to reduce water and increase oil production. The evaluation parameters for flow field adjustment are determined to be two indicator parameters: monthly water cut increase rate of a single well and daily oil-water change of a single well. The data distribution characteristics of the evaluation indicator parameters of the target reservoir are statistically analyzed to determine the evaluation standards, and the samples are labeled according to the standards.

6. The method for determining the main controlling factors of flow field regulation in old water flooding reservoirs based on data-driven according to claim 1 is characterized in that: In step 5, the characteristic parameters describing the flow field have multi-dimensional and multi-valued characteristics, and the data are complex and diverse. The continuous data of different attributes are normalized or discretized accordingly in combination with the physical meaning of the data to form sample data.

7. The method for determining the main controlling factors of flow field regulation in old water flooding reservoirs based on data-driven according to claim 1 is characterized in that: In step 6, the association analysis method based on the Apriori algorithm is used to mine the association rules that lead to the effectiveness of flow field regulation, and the main control parameters are determined according to the support and confidence of the rules.

Images