Well group injection and production scheme optimization method, device, equipment and medium based on logistic regression algorithm

Abstract

The present invention discloses a method, device, equipment and medium for optimizing a well group injection and production scheme based on a logistic regression algorithm, including: dividing a training set and a test set; constructing a sliding time window optimization algorithm based on a Pearson correlation coefficient algorithm; constructing a production prediction model based on logistic regression, and training the production prediction model at varying time intervals and a number of different preset time periods; verifying the production prediction model based on a test set; determining the upper limit of water injection and the upper limit of liquid production of the well group; taking the upper limit of water injection and the upper limit of liquid production as extreme values, traversing a number of combinations of water injection and liquid production, and predicting the unit time oil production of each combination based on the production prediction model; taking the maximization of unit time oil production and the minimum adjustment degree of the combination as the goal, outputting the optimal combination of water injection and liquid production of the well group. The present invention belongs to the field of well network oil production. The present invention can improve the development and management efficiency of water-injected oil fields.

Description

Technical Field

[0001] The present invention relates to the field of well network oil production, and in particular to a method, device, equipment and medium for optimizing a well group injection and production plan based on a logistic regression algorithm. Background Art

[0002] Waterflooding is a technique used to increase oil field recovery during oil extraction. When an oil field begins production, the natural pressure in the formation helps push the oil toward the wellbore. But over time, as oil is continuously extracted, the formation pressure gradually decreases, which can lead to a decrease in oil production. In order to maintain or restore the formation pressure and drive more oil to the production well, water (and sometimes gas or other fluids) can be injected into the reservoir at specific locations.

[0003] At present, most domestic oil fields have many production wells, complex processes, and are gradually entering the medium and high water content stage. The medium and high water content stage is an important window for water injection management. If the management is unreasonable, the comprehensive water content and natural decline rate will increase, and even seriously affect the final recovery rate of the oil field. Therefore, how to adjust the injection and production plan to make the oil production income higher is an urgent problem to be solved. Summary of the invention

[0004] The present invention solves the technical problem of the prior art that the injection and production scheme is unreasonable in the medium and high water content stages, resulting in low oil production returns, by providing a well group injection and production scheme optimization method, device, equipment and medium based on a logistic regression algorithm, thereby achieving the technical effect of providing a more reasonable injection and production scheme.

[0005] In a first aspect, the present invention provides a method for optimizing a well group injection-production plan based on a logistic regression algorithm, the method comprising:

[0006] S11, obtaining water injection well characteristic data and oil production well characteristic data related to the oil production per unit time of the well group, and dividing the training set and the test set based on the water injection well characteristic data and the oil production well characteristic data;

[0007] S12, constructing a sliding time window optimization algorithm based on the Pearson correlation coefficient algorithm, the sliding time window optimization algorithm is used to obtain the change time interval between the water injection wells and the oil production wells in the well group;

[0008] S13, constructing a yield prediction model based on logistic regression, and training the yield prediction model at varying time intervals and a plurality of different preset time periods;

[0009] S14, verifying the yield prediction model based on the test set, and outputting the yield prediction model when the yield prediction model meets the preset requirements;

[0010] S15, determining the upper limit of water injection volume and liquid production volume of the well group;

[0011] S16, taking the upper limit of water injection rate and the upper limit of liquid production rate as extreme values, traversing several combinations of water injection rate and liquid production rate, and predicting the oil production per unit time of each combination based on the production prediction model, wherein one combination includes a water injection rate and a liquid production rate;

[0012] S17, with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination, output the optimal combination of water injection volume and liquid production volume of the well group.

[0013] Furthermore, with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination, the optimal combination of water injection rate and liquid production rate of the well group is output, including:

[0014] Taking the maximization of oil production per unit time as the goal, several combinations of water injection volume and liquid production volume with the highest oil production were screened out;

[0015] Taking the minimum adjustment degree of the combination as the goal, from several combinations of water injection rate and liquid production rate with the largest oil production, select the combination of water injection rate and liquid production rate with the minimum adjustment degree;

[0016] The combination of water injection rate and liquid production rate with the smallest adjustment degree is taken as the optimal combination of water injection rate and liquid production rate of the well group.

[0017] Furthermore, the method further comprises:

[0018] Update the water injection well characteristic data and oil production well characteristic data of the well group;

[0019] When the characteristic data of the water injection wells and the characteristic data of the oil production wells of the well group continue to change and meet the preset production time, steps S11-S17 are re-executed.

[0020] Furthermore, the upper limit of water injection volume of the well group is determined, including:

[0021] Determine the maximum allowable injection pressure of the injection wells based on the geological conditions of the well group and the pressure resistance of the equipment of the injection wells in the well group;

[0022] Determine the upper limit of water injection volume based on the maximum allowable water injection pressure and water absorption index.

[0023] Furthermore, the upper limit of liquid production is determined, including:

[0024] Determine the liquid production capacity of the oil wells according to the geological conditions of the well group;

[0025] Determine the performance parameters of the lifting equipment under the production mode of the oil well;

[0026] The smaller value between the liquid production capacity corresponding to the performance parameters of the lifting equipment and the liquid production capacity of the oil well is used as the upper limit of the liquid production.

[0027] Furthermore, the production prediction model is trained at varying time intervals and a number of different preset time periods, including:

[0028] Inputting the characteristic data of the water injection well and the characteristic data of the oil production well into the production prediction model, predicting the predicted oil production of the well group in the preset time period after the change time interval, wherein a total of predicted oil production corresponding to a plurality of preset time periods is predicted;

[0029] The production prediction model is trained according to a number of predicted oil productions and the actual oil productions corresponding to each preset time period, wherein the loss function of the production prediction model training is MAE;

[0030] Based on the optimizer Adam, the neural network parameters of the yield prediction model are updated;

[0031] When the MAE reaches the preset application requirements, the yield prediction model is output.

[0032] Furthermore, the characteristic data of water injection wells and oil production wells related to the oil production per unit time of the well group are obtained, and the training set and the test set are divided based on the characteristic data of water injection wells and the characteristic data of oil production wells, including:

[0033] Obtain historical production data of well groups;

[0034] The historical production data are screened to obtain the characteristic data of water injection wells and oil production wells related to the oil production per unit time of the well group. The characteristic data of water injection wells include: daily water injection volume, wellhead water injection pressure, cumulative water injection volume, cumulative water injection time, layered injection ratio, apparent water absorption index, water injection layer thickness, and permeability. The characteristic data of oil production wells include: oil pressure, casing pressure, daily oil production, water content, nozzle diameter, production mode, cumulative production time, cumulative oil production, cumulative water production, cumulative liquid production, and permeability.

[0035] The characteristic data of water injection wells and oil production wells are spliced ​​to obtain wide data tables in several dimensions;

[0036] Determine the logic of data usage based on the correlation between the wide data table in several dimensions and the oil production per unit time of the well group;

[0037] Based on the data usage logic, the training set and test set are divided.

[0038] In a second aspect, the present invention provides a well group injection and production scheme optimization device based on a logistic regression algorithm, the device comprising:

[0039] An acquisition module is used to acquire the characteristic data of water injection wells and oil production wells related to the oil production per unit time of the well group, and divide the training set and the test set based on the characteristic data of water injection wells and the characteristic data of oil production wells;

[0040] A time interval determination module is used to construct a sliding time window optimization algorithm based on the Pearson correlation coefficient algorithm, and the sliding time window optimization algorithm is used to obtain the changing time interval between the water injection wells and the oil production wells in the well group;

[0041] A model training module is used to build a yield prediction model based on logistic regression and train the yield prediction model at varying time intervals and several different preset time periods;

[0042] The model verification module is used to retrain the yield prediction model based on the training set and the test set, and output the yield prediction model when the yield prediction model meets the preset requirements;

[0043] An upper limit determination module is used to determine the upper limit of water injection volume and liquid production volume of the well group;

[0044] The maximum prediction module is used to traverse several combinations of water injection rate and liquid production rate with the upper limit of water injection rate and liquid production rate as extreme values, and predict the oil production per unit time of each combination based on the production prediction model, wherein one combination includes one water injection rate and one liquid production rate;

[0045] The minimum adjustment module is used to output the optimal combination of water injection volume and liquid production volume of the well group with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination.

[0046] In a third aspect, the present invention provides an electronic device, comprising:

[0047] processor;

[0048] a memory for storing processor-executable instructions;

[0049] The processor is configured to execute to implement the well group injection and production plan optimization method based on the logistic regression algorithm as in the first aspect.

[0050] In a fourth aspect, the present invention provides a non-temporary computer-readable storage medium. When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to implement the well group injection and production plan optimization method based on the logistic regression algorithm as provided in the first aspect.

[0051] One or more technical solutions provided in the present invention have at least the following technical effects or advantages:

[0052] The present invention provides a method for optimizing a well group injection and production scheme based on a logistic regression algorithm, the method comprising: obtaining characteristic data of water injection wells and oil production wells related to the oil production per unit time of the well group, and dividing a training set and a test set based on the characteristic data of water injection wells and oil production wells; constructing a sliding time window optimization algorithm based on a Pearson correlation coefficient algorithm, the sliding time window optimization algorithm is used to obtain the changing time interval between water injection wells and oil production wells in the well group; constructing a production prediction model based on logistic regression, and optimizing the production with the changing time interval and a number of different preset time periods. The production prediction model is trained; the production prediction model is verified based on the test set, and when the production prediction model meets the preset requirements, the production prediction model is output; the upper limit of water injection volume and the upper limit of liquid production of the well group are determined; with the upper limit of water injection volume and the upper limit of liquid production as extreme values, several combinations of water injection volume and liquid production are traversed, and the unit time oil production of each combination is predicted based on the production prediction model, wherein one combination includes a water injection volume and a liquid production volume; with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination, the optimal combination of water injection volume and liquid production of the well group is output.

[0053] The present invention makes decisions through well group production composition analysis and injection and production parameter optimization; wherein, the well group production composition analysis is based on the extraction of characteristic variables related to the well group production, and the extracted historical characteristic information is used to construct the logical relationship between each characteristic variable and the production through a logistic regression algorithm to clarify the influence of each characteristic variable on the well group production. The injection and production parameter optimization decision is based on the production composition analysis, and further analyzes the influence characteristics of each parameter on the recovery rate at different production stages; by effectively controlling the injection and production parameters of each layer of the injection well at each production stage, the injection and production balance is achieved to reduce the fingering of the injected water and improve the sweep efficiency of the injected water, thereby effectively realizing the improvement of the recovery rate; and based on historical state learning, the optimal injection and production system scheme for improving the recovery rate under the current production conditions is recommended.

[0054] The present invention aims to realize online, near real-time, and dynamic formulation of injection and production work through data-driven methods, effectively solve the contradiction caused by the increasing demand for refined management of water-injected oil fields through conventional analysis methods, improve the efficiency of water-injected oil field development and management, improve the effect of water-injection oil displacement, and help improve the final recovery rate of oil fields. The present invention also clarifies that in the process of optimizing the injection and production scheme of multi-layer production wells, it is necessary to effectively control the injection and production parameters of each layer of the injection well, reduce the quality of the injected water through the injection and production balance, and improve the sweep efficiency of the injected water, so as to effectively realize the improvement of the recovery rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0056] Figure 1 A flow chart of a method for optimizing a well group injection-production scheme based on a logistic regression algorithm provided by the present invention;

[0057] Figure 2 A schematic flow chart of another method for optimizing a well group injection-production scheme based on a logistic regression algorithm provided by the present invention. DETAILED DESCRIPTION

[0058] The embodiment of the present invention solves the technical problem in the prior art that the injection and production scheme is unreasonable in the medium-high water content stage, resulting in low oil production income, by providing a method for optimizing the injection and production scheme of a well group based on a logistic regression algorithm.

[0059] The technical solution of the present invention is to solve the above technical problems, and the overall idea is as follows:

[0060] A method for optimizing a well group injection and production plan based on a logistic regression algorithm, the method comprising: S11, obtaining water injection well characteristic data and oil production well characteristic data related to the oil production per unit time of the well group, and dividing a training set and a test set based on the water injection well characteristic data and the oil production well characteristic data; S12, constructing a sliding time window optimization algorithm based on a Pearson correlation coefficient algorithm, the sliding time window optimization algorithm is used to obtain the changing time interval between the water injection well and the oil production well in the well group; S13, constructing a production prediction model based on logistic regression, and optimizing the production prediction model with the changing time interval and a number of different preset time periods Conduct training; S14, verify the production prediction model based on the test set, and output the production prediction model when it meets the preset requirements; S15, determine the upper limit of water injection volume and the upper limit of liquid production of the well group; S16, take the upper limit of water injection volume and the upper limit of liquid production as extreme values, traverse several combinations of water injection volume and liquid production, and predict the oil production per unit time of each combination based on the production prediction model, one of which contains a water injection volume and a liquid production volume; S17, with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination, output the optimal combination of water injection volume and liquid production of the well group.

[0061] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0062] First of all, the term "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0063] The present invention provides Figure 1 The well group injection and production scheme optimization method based on the logistic regression algorithm shown includes steps S11-S17:

[0064] The well group injection and production scheme optimization method based on the logistic regression algorithm provided by the present invention can also be referred to Figure 2 .

[0065] Regarding S11, the water injection well characteristic data and the oil production well characteristic data related to the oil production per unit time of the well group are obtained, and the training set and the test set are divided based on the water injection well characteristic data and the oil production well characteristic data.

[0066] Specifically, it includes: obtaining the historical production data of the well group; screening the historical production data to obtain the water injection well characteristic data and the oil production well characteristic data related to the oil production per unit time of the well group, wherein the water injection well characteristic data include: daily water injection volume, wellhead water injection pressure, cumulative water injection volume, cumulative water injection time, stratified injection ratio, apparent water absorption index, water injection layer thickness, permeability, and the oil production well characteristic data include: oil pressure, casing pressure, daily oil production, water content, nozzle diameter, production mode, cumulative production time, cumulative oil production, cumulative water production, cumulative liquid production, and permeability; splicing the water injection well characteristic data and the oil production well characteristic data to obtain a wide data table under several dimensions; determining the data usage logic according to the correlation between the wide data table under several dimensions and the oil production per unit time of the well group; and dividing the training set and the test set based on the data usage logic.

[0067] A well group includes several oil wells and several water wells. The number of water wells and oil wells can be determined according to actual conditions. The production data of the well group includes static data of the reservoir, historical production data of water injection wells, and historical production data of oil production wells. Among them, static data of the reservoir refers to data that does not change significantly over time, such as reservoir thickness, permeability, and porosity.

[0068] The static data of oil reservoirs, historical production data of water injection wells, and historical production data of oil production wells can be screened to select data that can affect the oil production per unit time of the well group.

[0069] The screened data are divided into water injection well characteristic data and oil production well characteristic data. The specific coverage of the water injection well characteristic data and the oil production well characteristic data can be referred to the above content and will not be elaborated here.

[0070] The obtained water injection well characteristic data and oil production well characteristic data are spliced ​​according to the information dimension distribution of each data and the well group relationship to form a wide data table of each dimensional level. The data information dimension distribution refers to the range, type and organization of each feature or variable in the data set. For example, some data does not change with time (permeability), and some data changes with time (daily oil production). The well group relationship refers to the interaction and connection between oil wells and oil wells, between oil wells and water wells, and between water wells in the well group based on factors such as geographical location, geological structure, and production dynamics.

[0071] Perform correlation analysis on each characteristic variable in the data wide table (i.e., injection well characteristic data and oil production well characteristic data) and the target variable (i.e., oil production per unit time) (this can be done by calculating statistical indicators such as the Pearson correlation coefficient and the Spearman rank correlation coefficient), and determine the importance of each feature to the target variable (oil production per unit time) through statistical methods or machine learning techniques (such as correlation coefficient matrix, feature selection algorithm, etc.).

[0072] The data usage logic is constructed based on the correlation results. For example, if certain data is highly correlated with the oil production per unit time, it will be retained, otherwise it will be discarded.

[0073] After the water injection well characteristic data and the oil production well characteristic data are merged into a data wide table, the data in the data wide table are divided into a training set and a test set according to a preset ratio. The training set and the test set are used to train the production prediction model based on logistic regression in the following text.

[0074] Regarding S12, a sliding time window optimization algorithm is constructed based on the Pearson correlation coefficient algorithm. The sliding time window optimization algorithm is used to obtain the changing time interval between the water injection wells and the oil production wells in the well group.

[0075] It should be noted that in oil field development, water injection wells increase the production of oil wells by injecting water into the reservoir. This process is called water flooding. Since the transmission of fluids in the formation takes time, there is a lag effect from water injection wells to oil wells. The time intervals for different oil wells to be affected by water injection may vary, depending on many factors, such as formation permeability, porosity, and distance between wells.

[0076] In order to determine the time delay corresponding to the lag effect (that is, the "change time interval" mentioned above) and determine the optimal time window for production prediction, the present invention uses the Pearson correlation coefficient combined with a sliding time window optimization algorithm. The Pearson correlation coefficient combined with a sliding time window optimization algorithm can help identify the time interval that has the greatest impact on the oil production of the oil production well after the change of the water injection well, and use the time interval that has the greatest impact as the change time interval.

[0077] Specifically: define a sliding time window, for example, from 1 day to 30 days; record the Pearson correlation coefficient under each time window, and find the time window with the highest correlation, which is the time interval when the change of injection well parameters has the greatest impact on the change of oil production in the production well.

[0078] Regarding S13, a yield prediction model based on logistic regression is constructed, and the yield prediction model is trained at varying time intervals and a number of different preset time periods.

[0079] Specifically, it includes: inputting water injection well characteristic data and oil production well characteristic data (that is, data in the training set) into the production prediction model, predicting the predicted oil production of the well group in a preset time period after the change time interval, wherein a total of predicted oil production corresponding to several preset time periods is predicted; training the production prediction model according to the several predicted oil productions and the actual oil production corresponding to each preset time period, wherein the loss function of the production prediction model training is MAE; updating the neural network parameters of the production prediction model based on the optimizer Adam; and outputting the production prediction model when MAE reaches the preset application requirements.

[0080] From the above wide data table, data related to oil production per unit time, namely, characteristic data of water injection wells and characteristic data of oil production wells, are obtained.

[0081] Due to the existence of a change time interval, when the parameters (or data) of the water injection well change, the oil production well will change after the change time interval. In other words, the oil production after the change time interval will be predicted. The preset time period can be determined based on actual conditions, such as 10 days, 15 days, or 20 days.

[0082] For example, the change time interval is 10 days, the preset time period is 15 days, and when the parameters of the water injection well change, the predicted oil production of the oil production well from the 11th to the 26th day (a total of 15 days) is predicted.

[0083] It is understandable that the above historical production data includes the actual oil production from 11 to 26. By comparing the actual oil production with the predicted oil production, the error of the production prediction model can be determined. It is also understandable that one preset time period corresponds to one predicted oil production and one actual oil production.

[0084] The loss function for training the yield prediction model may be MAE, which is the average of the absolute differences between the predicted value and the actual value. When MAE is lower than a preset threshold, it is considered that the preset application requirements are met, and the yield prediction model can be output.

[0085] Regarding S14, the yield prediction model is verified based on the test set, and when the yield prediction model meets the preset requirements, the yield prediction model is output.

[0086] The test set mentioned above is input into the production prediction model that has been trained for the first time, and the oil production of the well group is predicted again, and the verification is performed based on the verification set. The preset requirements (or preset application requirements) can be an error threshold or a maximum number of iterations. When the error threshold is met or the maximum number of iterations is reached, the production prediction model is output.

[0087] Regarding S15, determine the upper limit of water injection volume and liquid production volume of the well group.

[0088] Determining the upper limit of water injection volume of a well group includes: determining the maximum allowable water injection pressure of the water injection well according to the geological conditions of the well group and the pressure resistance of the equipment of the water injection wells in the well group; determining the upper limit of water injection volume according to the maximum allowable water injection pressure and water absorption index. Among them, the water absorption index refers to the water injection volume of the water injection well under unit water injection pressure.

[0089] According to the geological conditions and the pressure resistance of the equipment, the maximum allowable injection pressure of the injection well is determined. The maximum injection volume under the maximum allowable injection pressure is calculated using the water absorption index. The calculated maximum injection volume is used as the upper limit for optimizing the injection parameters, that is, the upper limit of the injection volume of the well group.

[0090] Determining the upper limit of liquid production includes: determining the liquid production capacity of the oil wells according to the geological conditions of the well group; determining the performance parameters of the lifting equipment under the production mode of the oil wells; and taking the smaller value between the liquid production capacity corresponding to the performance parameters of the lifting equipment and the liquid production capacity of the oil wells as the upper limit of liquid production.

[0091] Lifting equipment refers to equipment used to lift the bottom of the well to the surface, such as pumping units, electric submersible pumps, etc. Its performance parameters include maximum displacement, maximum power, maximum head, etc.

[0092] Regarding S16, taking the upper limit of water injection rate and the upper limit of liquid production rate as extreme values, several combinations of water injection rate and liquid production rate are traversed, and the oil production per unit time of each combination is predicted based on the production prediction model, wherein one combination includes a water injection rate and a liquid production rate.

[0093] Taking the upper limit of water injection rate and liquid production rate as extreme values, the reinforcement learning method is used to construct several combinations of water injection rate and liquid production rate.

[0094] Reinforcement learning is a machine learning method that learns the optimal strategy through interaction with the environment, including: agent, environment, state, action (actions that the agent can take, such as adjusting water injection parameters or liquid production parameters), reward, etc.

[0095] Each combination is input into the production prediction model to determine the predicted oil production under each combination.

[0096] Regarding S17, with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination, the optimal combination of water injection rate and liquid production rate of the well group is output.

[0097] Specifically, it includes: taking maximization of oil production per unit time as the goal, screening out several combinations of water injection rate and liquid production with the largest oil production; taking minimum adjustment degree of the combination as the goal, from several combinations of water injection rate and liquid production with the largest oil production, screening out the combination of water injection rate and liquid production with the smallest adjustment degree; and taking the combination of water injection rate and liquid production with the smallest adjustment degree as the optimal combination of water injection rate and liquid production for the well group.

[0098] It is understandable that there may be multiple combinations of water injection and liquid production that have the maximum oil production per unit time. The combination with the smallest adjustment of water injection and liquid production can be taken as the optimal combination of water injection and liquid production.

[0099] For example, the predicted oil production of groups A, B, and C is the same, all 100 barrels / day, the original water injection volume is 50m³ / d, the water injection pressure is 10MPa; the original pumping rate is 100 barrels / day;

[0100] Combination A: water injection volume 60m³ / d; oil pumping rate 100 barrels / day;

[0101] Combination B: water injection volume 49m³ / d; oil pumping rate 100 barrels / day;

[0102] Combination C: water injection volume 47m³ / d; oil pumping rate 100 barrels / day;

[0103] Therefore, the adjustment amount of combination B is the smallest, and combination B can be selected as the optimal combination of water injection rate and liquid production rate.

[0104] The method also includes: updating the water injection well characteristic data and the oil production well characteristic data of the well group; when the water injection well characteristic data and the oil production well characteristic data of the well group continue to change and meet the preset production time, re-execute steps S11-S17.

[0105] In summary, the present invention provides a method for optimizing the injection and production scheme of a well group based on a logistic regression algorithm, the method comprising: obtaining characteristic data of water injection wells and oil production wells related to the oil production per unit time of the well group, and dividing a training set and a test set based on the characteristic data of water injection wells and oil production wells; constructing a sliding time window optimization algorithm based on the Pearson correlation coefficient algorithm, the sliding time window optimization algorithm is used to obtain the changing time interval between the water injection wells and the oil production wells in the well group; constructing a production prediction model based on logistic regression, and using the changing time interval and several different preset time periods to predict the production of the water injection wells and the oil production wells. , train the production prediction model; verify the production prediction model based on the test set, and output the production prediction model when the production prediction model meets the preset requirements; determine the upper limit of water injection volume and the upper limit of liquid production of the well group; take the upper limit of water injection volume and the upper limit of liquid production as extreme values, traverse several combinations of water injection volume and liquid production, and predict the unit time oil production of each combination based on the production prediction model, one of which contains a water injection volume and a liquid production volume; with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination, output the optimal combination of water injection volume and liquid production of the well group.

[0106] The present invention makes decisions through well group production composition analysis and injection and production parameter optimization; wherein, the well group production composition analysis is based on the extraction of characteristic variables related to the well group production, and the extracted historical characteristic information is used to construct the logical relationship between each characteristic variable and the production through a logistic regression algorithm to clarify the influence of each characteristic variable on the well group production. The injection and production parameter optimization decision is based on the production composition analysis, and further analyzes the influence characteristics of each parameter on the recovery rate at different production stages; by effectively controlling the injection and production parameters of each layer of the injection well at each production stage, the injection and production balance is achieved to reduce the fingering of the injected water and improve the sweep efficiency of the injected water, thereby effectively realizing the improvement of the recovery rate; and based on historical state learning, the optimal injection and production system scheme for improving the recovery rate under the current production conditions is recommended.

[0107] The present invention aims to realize online, near real-time, and dynamic formulation of injection and production work through data-driven methods, effectively solve the contradiction caused by the increasing demand for refined management of water-injected oil fields through conventional analysis methods, improve the efficiency of water-injected oil field development and management, improve the effect of water-injection oil displacement, and help improve the final recovery rate of oil fields. The present invention also clarifies that in the process of optimizing the injection and production scheme of multi-layer production wells, it is necessary to effectively control the injection and production parameters of each layer of the injection well, reduce the quality of the injected water through the injection and production balance, and improve the sweep efficiency of the injected water, so as to effectively realize the improvement of the recovery rate.

[0108] Based on the same inventive concept, the present invention also provides a well group injection and production plan optimization device based on a logistic regression algorithm, the device comprising:

[0109] An acquisition module is used to acquire the characteristic data of water injection wells and oil production wells related to the oil production per unit time of the well group, and divide the training set and the test set based on the characteristic data of water injection wells and the characteristic data of oil production wells;

[0110] A time interval determination module is used to construct a sliding time window optimization algorithm based on the Pearson correlation coefficient algorithm, and the sliding time window optimization algorithm is used to obtain the changing time interval between the water injection wells and the oil production wells in the well group;

[0111] A model training module is used to build a yield prediction model based on logistic regression and train the yield prediction model at varying time intervals and several different preset time periods;

[0112] The model verification module is used to verify the yield prediction model based on the test set, and output the yield prediction model when the yield prediction model meets the preset requirements;

[0113] An upper limit determination module is used to determine the upper limit of water injection volume and liquid production volume of the well group;

[0114] The maximum prediction module is used to traverse several combinations of water injection rate and liquid production rate with the upper limit of water injection rate and liquid production rate as extreme values, and predict the oil production per unit time of each combination based on the production prediction model, wherein one combination includes one water injection rate and one liquid production rate;

[0115] The minimum adjustment module is used to output the optimal combination of water injection volume and liquid production volume of the well group with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination.

[0116] Based on the same inventive concept, the present invention also provides an electronic device as shown, including:

[0117] processor;

[0118] a memory for storing processor-executable instructions;

[0119] The processor is configured to execute to implement the well group injection and production plan optimization method based on the logistic regression algorithm as provided above.

[0120] Based on the same inventive concept, the present invention also provides a non-temporary computer-readable storage medium. When the instructions in the storage medium are executed by the processor of an electronic device, the electronic device can execute the well group injection and production plan optimization method based on the logistic regression algorithm as provided above.

[0121] Since the electronic device introduced in this embodiment is an electronic device used to implement the information processing method in the embodiment of the present invention, based on the information processing method introduced in the embodiment of the present invention, a person skilled in the art can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of the present invention is not described in detail here. As long as the electronic device used by a person skilled in the art to implement the information processing method in the embodiment of the present invention, it belongs to the scope of protection of the present invention.

[0122] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0123] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0124] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0125] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1The steps for the functions specified in one or more boxes.

[0126] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0127] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for optimizing well group injection and production scheme based on logistic regression algorithm, characterized in that: The method comprises: S11, acquiring water injection well characteristic data and oil production well characteristic data related to the oil production per unit time of the well group, and dividing the training set and the test set based on the water injection well characteristic data and the oil production well characteristic data; S12, constructing a sliding time window optimization algorithm based on the Pearson correlation coefficient algorithm, wherein the sliding time window optimization algorithm is used to obtain the change time interval between the water injection wells and the oil production wells in the well group; S13, constructing a yield prediction model based on logistic regression, and training the yield prediction model at the change time interval and a plurality of different preset time periods; S14, verifying the yield prediction model based on the test set, and outputting the yield prediction model when the yield prediction model meets the preset requirements; S15, determining the upper limit of water injection volume and liquid production volume of the well group; S16, taking the upper limit of the water injection rate and the upper limit of the liquid production rate as extreme values, traversing several combinations of water injection rates and liquid production rates, and predicting the oil production per unit time of each combination based on the production prediction model, wherein one combination includes a water injection rate and a liquid production rate; S17, with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination, output the optimal combination of water injection volume and liquid production volume of the well group.

2. The method for optimizing the well group injection and production plan based on the logistic regression algorithm according to claim 1, characterized in that: With the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination, the optimal combination of water injection and liquid production of the well group is output, including: Taking the maximization of oil production per unit time as the goal, several combinations of water injection volume and liquid production volume with the highest oil production were screened out; Taking the minimum adjustment degree of the combination as the goal, from several combinations of water injection rate and liquid production rate with the maximum oil production, select the combination of water injection rate and liquid production rate with the minimum adjustment degree; The combination of water injection rate and liquid production rate with the smallest adjustment degree is taken as the optimal combination of water injection rate and liquid production rate of the well group.

3. The method for optimizing the well group injection and production plan based on the logistic regression algorithm according to claim 1, characterized in that: The method further comprises: Update the water injection well characteristic data and oil production well characteristic data of the well group; When the characteristic data of the water injection wells and the characteristic data of the oil production wells of the well group continue to change and meet the preset production time, steps S11-S17 are re-executed.

4. The method for optimizing the well group injection and production plan based on the logistic regression algorithm according to claim 1, characterized in that: Determine the upper limit of water injection volume for the well group, including: Determine the maximum allowable injection pressure of the injection wells based on the geological conditions of the well group and the pressure resistance of the equipment of the injection wells in the well group; The upper limit of the water injection volume is determined according to the maximum allowable water injection pressure and the water absorption index.

5. The method for optimizing well group injection and production scheme based on logistic regression algorithm according to claim 1, characterized in that: Determine the upper limit of liquid production, including: Determine the liquid production capacity of the oil wells according to the geological conditions of the well group; Determine the performance parameters of the lifting equipment under the production mode of the oil well; The smaller value between the liquid production capacity corresponding to the performance parameter of the lifting equipment and the liquid production capacity of the oil well is used as the upper limit of the liquid production.

6. The method for optimizing well group injection and production scheme based on logistic regression algorithm according to claim 1, characterized in that: The production prediction model is trained at the changing time interval and a plurality of different preset time periods, including: Inputting the water injection well characteristic data and the oil production well characteristic data into the production prediction model, predicting the predicted oil production of the well group in a preset time period after the change time interval, wherein a total of predicted oil production corresponding to a plurality of preset time periods is predicted; The production prediction model is trained according to a number of predicted oil productions and actual oil productions corresponding to each preset time period, wherein the loss function of the production prediction model training is MAE; Based on the optimizer Adam, the neural network parameters of the yield prediction model are updated; When the MAE reaches the preset application requirement, the yield prediction model is output.

7. The method for optimizing well group injection and production scheme based on logistic regression algorithm according to claim 1, characterized in that: Acquiring water injection well characteristic data and oil production well characteristic data related to the oil production per unit time of the well group, and dividing the training set and the test set based on the water injection well characteristic data and the oil production well characteristic data, including: Obtain historical production data of well groups; The historical production data are screened to obtain water injection well characteristic data and oil production well characteristic data related to the oil production per unit time of the well group, wherein the water injection well characteristic data include: daily water injection volume, wellhead water injection pressure, cumulative water injection volume, cumulative water injection time, layered injection ratio, apparent water absorption index, water injection layer thickness, and permeability; the oil production well characteristic data include: oil pressure, casing pressure, daily oil production, water content, nozzle diameter, production mode, cumulative production time, cumulative oil production, cumulative water production, cumulative liquid production, and permeability; The characteristic data of water injection wells and oil production wells are spliced ​​to obtain wide data tables in several dimensions; Determine the logic of data usage based on the correlation between the wide data table in several dimensions and the oil production per unit time of the well group; Based on the data usage logic, the training set and the test set are divided.

8. A well group injection and production plan optimization device based on a logistic regression algorithm, characterized in that: The device comprises: An acquisition module, used to acquire water injection well characteristic data and oil production well characteristic data related to the oil production per unit time of the well group, and divide the training set and the test set based on the water injection well characteristic data and the oil production well characteristic data; A time interval determination module, used to construct a sliding time window optimization algorithm based on a Pearson correlation coefficient algorithm, wherein the sliding time window optimization algorithm is used to obtain a change time interval between a water injection well and an oil production well in a well group; A model training module, used for constructing a yield prediction model based on logistic regression, and training the yield prediction model at the change time interval and a number of different preset time periods; A model verification module, used to verify the yield prediction model based on the test set, and output the yield prediction model when the yield prediction model meets the preset requirements; An upper limit determination module is used to determine the upper limit of water injection volume and liquid production volume of the well group; a maximum prediction module, for taking the upper limit of the water injection rate and the upper limit of the liquid production rate as extreme values, traversing several combinations of water injection rates and liquid production rates, and predicting the oil production per unit time of each combination based on the production prediction model, wherein one combination includes a water injection rate and a liquid production rate; The minimum adjustment module is used to output the optimal combination of water injection volume and liquid production volume of the well group with the goal of maximizing the oil production per unit time and minimizing the adjustment degree of the combination.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to execute and implement the well group injection and production plan optimization method based on the logistic regression algorithm as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that: When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device is enabled to implement the well group injection and production plan optimization method based on the logistic regression algorithm as described in any one of claims 1 to 7.

Images