A method and device for predicting oil field output

Abstract

The invention discloses a method and device for predicting oil field production, and relates to the technical field of machine learning. The method uses a machine learning algorithm to automatically find a better model suitable for predicting the production of a target oil field from a plurality of known oil field production prediction models, and also ensures the accuracy of predicting the oil field production. The main technical scheme of the invention is: receiving an oil field production prediction instruction, the prediction instruction carries a target oil field identifier and current production data of the target oil field; according to the target oil field identifier, obtaining historical production data corresponding to the target oil field; obtaining multiple oil field production prediction models, each oil field production prediction model is encapsulated into an independent operator; using the historical production data to screen multiple operators, and screening out an operator matching the target oil field from the multiple operators; processing the current production data by calling the operator matching the target oil field, executing an operation of predicting the future production of the target oil field, and outputting a prediction result.

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning, and in particular to a method and device for predicting oil field output. Background Art

[0002] The change of oil field production is very important to an oil field, and it is an important factor affecting the oil field development and production process and investment decision-making. Accurately predicting the decline law of oil field production and the remaining oil is very important for oil field development, and it is the basis for scientific management of oil fields and the formulation of economically feasible development plans. If the oil field production can be accurately predicted, the oil field that has been put into production can be better managed, the oil field production can be reasonably controlled and optimized, more oil can be extracted from the ground, and the recovery rate and development effect of the oil field can be improved.

[0003] At present, each oil field in China has its own oil field production management system. The daily oil production, water production, and liquid production data of each oil field, each oil production plant, and each well are recorded in the production management system library. Therefore, the oil fields have accumulated a large amount of production data. Based on these production data, hundreds of oil field production prediction technologies and methods have been proposed since the 20th century.

[0004] However, although there are many methods for predicting oil field production, each method has its limitations, and no method or technology can be applied to accurately predict the production of all oil fields. Now, after determining the target oil field to be tested, it is necessary to manually find a prediction method and model suitable for the target oil field from a large number of known prediction methods, which is not only difficult and labor-intensive, but also complex in calculation and difficult to implement. Summary of the invention

[0005] In view of this, the present invention provides a method and device for predicting oil field production, the main purpose of which is to use a machine learning algorithm to automatically find a better model suitable for predicting the target oil field production from a number of known oil field production prediction models, so as to use the better model to predict the future production of the oil field to ensure the accuracy of the prediction results.

[0006] In order to achieve the above object, the present invention mainly provides the following technical solutions:

[0007] In one aspect, the present invention provides a method for predicting oil field production, the method comprising:

[0008] receiving an oilfield production prediction instruction, wherein the prediction instruction carries a target oilfield identifier and current production data of the target oilfield;

[0009] According to the target oil field identifier, obtaining the historical production data corresponding to the target oil field;

[0010] Acquire multiple oilfield production prediction models, each of which is encapsulated into an independent operator;

[0011] Using the historical production data to screen the multiple operators, and selecting an operator matching the target oil field from the multiple operators;

[0012] The current production data is processed by calling an operator matching the target oil field, an operation of predicting the future production of the target oil field is performed, and a prediction result is output.

[0013] Optionally, the using the historical production data to screen the multiple operators, and selecting an operator matching the target oil field from the multiple operators, includes:

[0014] Using historical production data to train multiple operators respectively, and obtaining the optimal parameters corresponding to each operator;

[0015] According to the optimal parameters, the corresponding operator is adjusted to obtain an operator that matches the target oil field production.

[0016] Optionally, after selecting an operator matching the target oil field from the multiple operators, if there are multiple operators matching the target oil field, the method further includes:

[0017] If there are multiple operators matching the target oil field, the operators matching the target oil field are determined as operators to be selected;

[0018] According to a preset time unit, the historical production data is divided into partial production data corresponding to each preset time unit;

[0019] Using each of the operators to be selected to process the part of the production data, to obtain a prediction result corresponding to each of the time units;

[0020] Find the historical actual results corresponding to each of the time units;

[0021] Calculate the average error percentage of each of the to-be-selected operators in processing the historical production data according to the prediction results and historical actual results corresponding to each of the time units;

[0022] The operator whose average error percentage meets the preset conditions is selected as the target operator, and the preset conditions are the top N operators with the smallest average error percentage, or the top M% operators with the smallest average error percentage, where N is a positive integer and M is a positive number.

[0023] Optionally, the processing of the current production data by calling an operator matching the target oil field, performing an operation of predicting the future production of the target oil field, and outputting a prediction result includes:

[0024] Calling the target operator to process the current production data, and outputting the prediction results corresponding to each target operator;

[0025] The prediction results corresponding to each of the target operators are respectively combined into a prediction result set;

[0026] The prediction result set is fed back to the user.

[0027] Optionally, before screening the plurality of operators by using the historical production data to select an operator matching the target oil field from the plurality of operators, the method further includes:

[0028] Analyzing the parameter dimensions included in the historical production data;

[0029] By comparing the parameter dimension with the parameter dimension that the operator needs to process, determining whether they match;

[0030] If there is no match, the operator is discarded.

[0031] Optionally, before screening the plurality of operators by using the historical production data to select an operator matching the target oil field from the plurality of operators, the method further includes:

[0032] By adapting the historical production data to each of the operators, outputting the prediction result of each of the operators on the future production of the target oil field;

[0033] According to the time interval corresponding to the prediction result, obtaining the historical actual result corresponding to the target oil field in the same time interval;

[0034] By comparing the prediction result calculated by each operator with the historical actual result, it is determined whether the prediction error reaches a preset threshold;

[0035] If so, the operator is discarded.

[0036] On the other hand, the present invention also provides a device for predicting oil field production, the device comprising:

[0037] A receiving unit, used for receiving an oilfield production prediction instruction, wherein the prediction instruction carries a target oilfield identifier and current production data of the target oilfield;

[0038] an acquisition unit, configured to search for historical production data corresponding to the target oil field according to the target oil field identifier received by the receiving unit;

[0039] The acquisition unit is further used to acquire multiple oilfield production prediction models, each of which is encapsulated into an independent operator;

[0040] a screening unit, configured to screen the operators acquired by the acquisition unit using the historical production data received by the receiving unit, and select an operator matching the target oil field from the multiple operators;

[0041] The execution unit is used to process the current production data by calling the operator matched with the target oil field screened by the screening unit, execute the operation of predicting the future production of the target oil field, and output the prediction result.

[0042] Optionally, the screening unit includes:

[0043] A training module, used to train multiple operators respectively using historical production data to obtain optimal parameters corresponding to each operator;

[0044] The determination module is used to adjust the corresponding operator according to the optimal parameters obtained by the training module, and determine the operator that matches the output of the target oil field.

[0045] Optionally, the device further comprises:

[0046] The determination unit is further configured to, after selecting an operator matching the target oil field from the plurality of operators, determine the operator matching the target oil field as the operator to be selected if there are multiple operators matching the target oil field;

[0047] A division unit, used for dividing the historical production data into partial production data corresponding to each preset time unit according to the preset time unit;

[0048] A processing unit, used to process the part of the production data using each of the operators to be selected to obtain a prediction result corresponding to each of the time units;

[0049] A search unit, used to obtain the historical actual result corresponding to each of the time units;

[0050] A calculation unit, configured to calculate an average error percentage obtained by each of the to-be-selected operators in processing the historical production data according to the prediction result obtained by the processing unit corresponding to each of the time units and the historical actual result obtained by the obtaining unit;

[0051] A selection unit is used to select an operator whose average error percentage calculated by the calculation unit meets a preset condition as a target operator, wherein the preset condition is the top N operators with the smallest average error percentage, or the top M% operators with the smallest average error percentage, wherein N is a positive integer and M is a positive number.

[0052] Optionally, the execution unit includes:

[0053] A calling module, used for calling the target operator to process the current production data;

[0054] An output module, used to output the prediction results corresponding to each of the target operators;

[0055] A composition module, used for composing the prediction results corresponding to each of the target operators into a prediction result set;

[0056] The feedback module is used to feed back the prediction result set to the user.

[0057] Optionally, the device further comprises:

[0058] A parsing unit, configured to parse the parameter dimensions contained in the historical production data before screening the plurality of operators using the historical production data to select an operator matching the target oil field from the plurality of operators;

[0059] A judging unit, configured to compare the parameter dimension obtained by the parsing unit with the parameter dimension to be processed by the operator to judge whether there is a match;

[0060] The discarding unit is used to discard the operator when the judging unit judges that the parameter dimension obtained by the parsing unit matches the parameter dimension that the operator needs to process.

[0061] Optionally, the device further comprises:

[0062] A prediction unit, used for outputting a prediction result of each operator on the future production of the target oil field by adapting the historical production data to each operator before screening the plurality of operators using the historical production data to screen out an operator matching the target oil field from the plurality of operators;

[0063] The search unit is further used to obtain the historical actual results corresponding to the target oil field in the same time interval according to the time interval corresponding to the prediction result;

[0064] The judging unit is further configured to judge whether a prediction error reaches a preset threshold by comparing the prediction result predicted by each operator with the historical actual result acquired by the acquiring unit;

[0065] The discarding unit is further configured to discard the operator when the judging unit judges that the prediction error reaches a preset threshold.

[0066] On the other hand, the present invention further provides a storage medium, which includes a stored program, wherein when the program is executed, the device where the storage medium is located is controlled to execute the above-mentioned method for predicting oil field production.

[0067] In another aspect, the present invention further provides an electronic device, the device comprising at least one processor, and at least one memory and a bus connected to the processor;

[0068] Wherein, the processor and the memory communicate with each other through the bus;

[0069] The processor is used to call the program instructions in the memory to execute the above-mentioned oil field production prediction method.

[0070] By means of the above technical solution, the technical solution provided by the present invention has at least the following advantages:

[0071] The present invention provides a method and device for predicting oil field production. The present invention uses a machine learning algorithm to automatically find a better model suitable for the target oil field production prediction from a number of known oil field production prediction models. Compared with the prior art, the present invention solves the problem of difficulty and workload, complex calculations, and practical difficulties caused by manually finding a prediction method and model suitable for the target oil field. The method provided by the present invention, for each target oil field to be tested, when receiving the current production data of the target oil field uploaded by the user, automatically and efficiently finds the better model for executing the prediction of the oil field production, and calls this better model to predict the future production of the target oil field, thereby ensuring the accuracy of predicting the oil field production.

[0072] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0074] Figure 1 A flow chart of a method for predicting oil field production provided by an embodiment of the present invention;

[0075] Figure 2 A flow chart of another method for predicting oil field production provided by an embodiment of the present invention;

[0076] Figure 3 A block diagram of a device for predicting oil field production provided by an embodiment of the present invention;

[0077] Figure 4 A block diagram of another oilfield production prediction device provided by an embodiment of the present invention;

[0078] Figure 5 An electronic device for predicting oil field production provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0079] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0080] The embodiment of the present invention provides a method for predicting oil field production, such as Figure 1 As shown, the method uses a machine learning algorithm to automatically find a better model suitable for the target oil field production prediction from a number of known oil field production prediction models. For this, the embodiment of the present invention provides the following specific steps:

[0081] 101. Receive an oil field production prediction instruction, where the prediction instruction carries a target oil field identifier and current production data of the target oil field.

[0082] The target oil field refers to the oil field for which the future production forecast is to be executed.

[0083] In an embodiment of the present invention, a human-computer interaction interface can be pre-constructed, and the interface is used to receive an instruction to predict the future production of a designated oil field. After parsing the target oil field identifier and the current production data of the target oil field carried in the prediction instruction, the background can directly call the relevant prediction method / model to automatically perform the prediction operation based on the data information carried by the prediction instruction. The automatically executed prediction operation is the content recorded in steps 102-105.

[0084] It should be noted that, in the embodiment of the present invention, it is equivalent to building a platform for predicting the future production of oil fields. For users, as long as a prediction instruction is issued and data information is specified, including: determining the target oil field to be predicted and inputting the current production data of the target oil field, the built platform will automatically execute the prediction and feedback the prediction results, so that the predicted oil field production operation is simple and efficient. This operation does not require users to have professional computer knowledge. Users can call optimized operators for production prediction without writing code, thereby lowering the threshold for using oil field production prediction.

[0085] 102. According to the target oil field identification, find the historical production data corresponding to the target oil field.

[0086] In the embodiment of the present invention, after the target oil field identifier is determined, historical production data can be found from the production management system corresponding to the target oil field.

[0087] It should be noted that the collected historical production data of the target oil field can be divided into a training set and a test set in a ratio of 8:2. If the historical production data is less than 5 points (i.e., data of 5 time units), such as 5 years or 5 days of historical production data, all data are divided into a training set. In the embodiment of the present invention, a training set is used to construct a machine learning model for finding an operator suitable for performing production prediction on the target oil field from multiple operators, and the test set is used to test whether the quality of the constructed machine learning model meets the standards.

[0088] 103. Obtain multiple oil field production prediction models, each of which is encapsulated into an independent operator.

[0089] At present, people have proposed hundreds of oil field production prediction technologies and methods. Generally speaking, oil field prediction methods can be divided into three categories: methods based on statistical models, methods based on mechanism models, and prediction methods based on system functions. Specific examples are as follows:

[0090] The method based on statistical model is to fit the historical oil and water production of oil fields, blocks or single wells from the perspective of mathematical statistics, and obtain formulas and models for various production changes to predict the future production of oil fields. There are many statistical prediction methods, mainly including water drive characteristic curve method, empirical formula method, Arps decline curve method, generalized Weng prediction model method, HCZ method, etc.

[0091] The method based on mechanism model mainly starts from the engineering mechanism of oil field production and uses technologies and methods such as reservoir numerical simulation, material balance, and seepage mechanics to predict the future output of the oil field.

[0092] The prediction method based on system function mainly predicts the output of oil fields from the perspective of systems theory and cybernetics.

[0093] In an embodiment of the present invention, the above-mentioned oil field production prediction method is encapsulated into an independent operator and stored in a machine learning library, so that after determining the target oil field to be predicted, a better operator suitable for predicting the future production of the target oil field is screened out from the machine learning library.

[0094] Furthermore, in the embodiment of the present invention, step 103 and step 101 are in parallel relationship, and there is no order restriction for these two steps.

[0095] 104. Use historical production data to screen multiple operators and select operators that match the target oil field from the multiple operators.

[0096] In an embodiment of the present invention, the historical production data are adapted to these operators one by one, and combined with the machine learning algorithms in the operators, all of which are regression algorithms, such as: HuberRegressor, Lasso, Ridge, SGDRegressor, LinearSVR, SVR, DecisionTreeRegressor, AdaBoostRegressor, BaggingRegressor, GradientBoostRegressor, RandomForestRegressor, and continuously adjust the parameters of a single operator so that the error of the operator predicting the oil field production after parameter adjustment is closer to the minimum direction and as close to the minimum error as possible, so as to obtain the operator after parameter adjustment, and then further select the better operator suitable for predicting the future production of the target oil field.

[0097] 105. Process the current production data by calling the operator matching the target oil field, execute the operation of predicting the future production of the target oil field, and output the prediction result.

[0098] In an embodiment of the present invention, after determining the optimal operator that matches the target oil field, according to step 101, in the platform that is built to execute the prediction of the future production of the oil field, the background calls the operator that matches the target oil field to process the current production data, execute the prediction operation, obtain the prediction result, and output the prediction result through the human-computer interaction interface to provide feedback to the user.

[0099] Alternatively, the platform can be combined to publish the best operator as an API through microservices for direct calls by business users. Users can select a specific operator and then perform the prediction operation according to the prediction instructions issued by the user. Regardless of the above methods, the prediction of future oil field production is simple, efficient, convenient and flexible.

[0100] The embodiment of the present invention provides a method for predicting oil field production. The embodiment of the present invention uses a machine learning algorithm to automatically find a better model suitable for the target oil field production prediction from a large number of known oil field production prediction models. Compared with the prior art, it solves the problem of manually finding a prediction method and model suitable for the target oil field from a large number of known prediction methods, which leads to high difficulty and workload, complex calculations, and practical difficulties. The method provided by the embodiment of the present invention, for each target oil field to be tested, when receiving the current production data of the target oil field uploaded by the user, automatically and efficiently finds the better model for executing the prediction of the oil field production, and calls this better model to predict the future production of the target oil field, and also ensures the accuracy of the predicted oil field production.

[0101] In order to explain the above embodiment in more detail, the embodiment of the present invention also provides another method for predicting oil field production, such as Figure 2 As shown, the method adds a pre-screening step for operators, thereby reducing the workload of subsequent machine learning to adjust operator parameters and select better operators. For this, the embodiment of the present invention provides the following specific steps:

[0102] 201. Receive an oil field production prediction instruction, where the prediction instruction carries a target oil field identifier and current production data of the target oil field.

[0103] 202. According to the target oil field identifier, search for the historical production data corresponding to the target oil field.

[0104] 203. Obtain multiple oil field production prediction models, each of which is encapsulated into an independent operator.

[0105] In the embodiment of the present invention, for the description of steps 201-203, please refer to steps 101-103, which will not be repeated here.

[0106] 204. Preliminarily filter out operators that are not suitable for predicting the future production of the target oil field from multiple operators.

[0107] In the embodiment of the present invention, a pre-screening step is added to perform on the operators, the purpose of which is to filter out operators that are obviously not suitable for predicting the future production of the target oil field, thereby reducing the workload of subsequent machine learning to adjust the operator parameters and select the better operator, and improving the efficiency of selecting the better operator. Specifically, the following two preliminary filtering methods can be included:

[0108] One preliminary filtering method is to preliminarily filter multiple existing operators based on training data, that is, data in historical production data used to train machine learning models. The specific steps can be as follows:

[0109] First, parse the parameter dimensions contained in the historical production data; second, compare the parameter dimensions with the parameter dimensions that the operator needs to process to determine whether they match. If not, the operator can be directly discarded.

[0110] For example, the parameter dimensions that the operator needs to process are: temperature, pressure, saturation, monthly oil production, and daily gas production. However, the historical production data only contains three parameter dimensions. Obviously, the data dimensions do not match, and the operator cannot be used to predict oil field production.

[0111] Another preliminary filtering method is to build a network based on the historical prediction results of the operators based on the Bayesian network method to filter out the operators whose prediction errors of the future production of the oil field are large and whose accuracy obviously does not meet the requirements. The specific statements are as follows:

[0112] By adapting the historical production data to each operator, the prediction results of each operator for the future production of the target oil field are output. According to the time interval corresponding to the prediction results, the historical actual results corresponding to the target oil field in the same time interval are obtained. By comparing the prediction results calculated by each operator with the historical actual results, it is determined whether the prediction error reaches the preset threshold. If so, the operator is discarded.

[0113] It should be noted that in the embodiment of the present invention, for the above two filtering methods, you can select one of them, or you can use them in combination. After preliminary screening of the encapsulated operators and removing those operators that are obviously not suitable for the target oil field production prediction, the remaining operators are modeled to form a set of machine learning and business mechanism models suitable for the current oil field production data.

[0114] 205. Use historical production data to train multiple operators respectively to obtain the optimal parameters corresponding to each operator. According to the optimal parameters, adjust the corresponding operator to obtain the operator that matches the target oil field production.

[0115] In the embodiment of the present invention, a machine learning algorithm is used to perform automatic parameter adjustment operations on each operator to obtain the optimal parameters corresponding to each operator. Specifically, the process of performing parameter adjustment on each operator may include: after each operator processes the historical production data to obtain the prediction result, it is compared with the historical actual result of the same time interval, and according to the difference between the prediction result and the historical actual result, the operator is automatically adjusted by machine learning, and the operator after parameter adjustment performs the prediction operation again, and so on and so forth until the difference between the prediction result and the historical actual result tends to be minimized. Therefore, the automatic parameter adjustment process is equivalent to the process in which the operator continuously performs the prediction operation and trial and error. Thus, the automatic parameter adjustment is used to replace the traditional manual parameter adjustment, and the system automatically adjusts the parameters of a single operator to automatically achieve the best prediction effect of a single operator.

[0116] Furthermore, for the embodiment of the present invention, if multiple operators matching the target oil field are obtained, a better operator can be further extracted from them to calculate the prediction result using the better operator, instead of feeding back the prediction result calculated by each matching operator to the user, so as to reduce the amount of feedback data and make the feedback prediction result more targeted, thereby improving the accuracy of the feedback prediction of the oil field production. Specifically, the steps of optimizing and selecting the better operator can include the following:

[0117] First, if there are multiple operators matching the target oil field, obtain the operators matching the target oil field as the operators to be selected. According to the preset time unit, divide the historical production data into partial production data corresponding to each preset time unit, use each operator to be selected to process the partial production data, and obtain the prediction result corresponding to each time unit.

[0118] For example, the historical production data contains multi-dimensional indicator data of the oil field for 5 years, so the preset time unit is selected as year, and the historical production data is divided into historical production data corresponding to each year.

[0119] Furthermore, the historical actual results corresponding to each time unit are obtained. For example, after determining that the preset time unit is a year, in the embodiment of the present invention, the historical actual results of each year are known data, so they can be directly obtained from the production management system corresponding to the current oil field.

[0120] Secondly, based on the prediction results and historical actual results corresponding to each time unit, the average error percentage obtained by each candidate operator processing historical production data is calculated.

[0121] Next, based on the average error percentage obtained by processing historical production data for each candidate operator, the operator whose average error percentage meets the preset conditions is selected as the target operator. The preset condition is the top N operators with the smallest average error percentage, or the top M% operators with the smallest average error percentage, where N is a positive integer and M is a positive number. The specific statements are as follows:

[0122] For example: sort the average error percentage corresponding to each operator to be selected in order from small to large percentage values ​​to obtain a value queue, and select a preset number of average error percentages in order from the first to the last in the value queue, that is, select the first N operators with the smallest average error percentage.

[0123] Another example: after obtaining the above-mentioned average error percentage numerical queue, according to the number of values ​​contained in the numerical queue, select the top M% operators with the smallest average error percentage, that is, select smaller values ​​from the numerical queue, but the selected number can be M% of the total number.

[0124] 206. Process the current production data by calling the operator matching the target oil field, execute the operation of predicting the future production of the target oil field, and output the prediction result.

[0125] In an embodiment of the present invention, after obtaining a better target operator based on step 205, the target operator is called to process the current production data, and the prediction results corresponding to each target operator are output. The prediction results corresponding to each target operator are combined into a prediction result set, and the prediction result set is pushed to the user.

[0126] It should be noted that, in the embodiment of the present invention, it is preferred to use multiple more optimal operators to predict the target oil field production so as to feedback multiple prediction results to the user. There will inevitably be differences in the prediction results, but this difference is already controllable and tends to the minimum error as much as possible. Since the prediction results are reference data for estimating the oil field production for the user and are not equal to the actual results in the future, by feeding back multiple prediction results to the user, it is equivalent to being able to feedback high, medium and low results to the user, so that the user can have a more comprehensive understanding of the analysis of the prediction of the future production of the oil field.

[0127] Furthermore, as a response to the above Figure 1 , Figure 2 The implementation of the method shown in the embodiment of the present invention provides an oil field production prediction device. The device embodiment corresponds to the aforementioned method embodiment. For ease of reading, the device embodiment will not repeat the details of the aforementioned method embodiment one by one, but it should be clear that the device in this embodiment can correspond to all the contents of the aforementioned method embodiment. The device is used to automatically find a better model suitable for the target oil field production prediction from many known oil field production prediction models, specifically as follows Figure 3 As shown, the device comprises:

[0128] The receiving unit 301 is used to receive an oilfield production prediction instruction, wherein the prediction instruction carries a target oilfield identifier and current production data of the target oilfield;

[0129] An acquiring unit 302 is configured to acquire historical production data corresponding to the target oil field according to the target oil field identifier received by the receiving unit 301;

[0130] The acquisition unit 302 is further used to acquire multiple oilfield production prediction models, each of which is encapsulated into an independent operator;

[0131] A screening unit 303 is used to screen the operators acquired by the acquisition unit 302 using the historical production data received by the receiving unit 301, and to select an operator matching the target oil field from the multiple operators;

[0132] The execution unit 304 is used to process the current production data by calling the operator matched with the target oil field screened by the screening unit 303, execute the operation of predicting the future production of the target oil field, and output the prediction result.

[0133] Further, such as Figure 4 As shown, the screening unit 303 includes:

[0134] The training module 3031 is used to train multiple operators respectively using historical production data to obtain the optimal parameters corresponding to each operator;

[0135] The determination module 3032 is used to adjust the corresponding operator according to the optimal parameters obtained by the training module 3031, and determine the operator that matches the output of the target oil field.

[0136] Further, such as Figure 4 As shown, the device also includes:

[0137] The determination unit 305 is further configured to, after selecting an operator matching the target oil field from the multiple operators, determine the operator matching the target oil field as the operator to be selected if there are multiple operators matching the target oil field;

[0138] A division unit 306, configured to divide the historical production data into partial production data corresponding to each preset time unit according to the preset time unit;

[0139] A processing unit 307 is used to process the part of the production data using each of the operators to be selected to obtain a prediction result corresponding to each of the time units;

[0140] A search unit 308, configured to search for the historical actual result corresponding to each of the time units;

[0141] A calculation unit 309 is used to calculate an average error percentage obtained by each of the to-be-selected operators processing the historical production data according to the prediction result obtained by the processing unit 307 corresponding to each of the time units and the historical actual results found by the search unit 308;

[0142] The selection unit 310 is used to select an operator whose average error percentage calculated by the calculation unit 309 meets a preset condition as a target operator, wherein the preset condition is the top N operators with the smallest average error percentage, or the top M% of operators with the smallest average error percentage, where N is a positive integer and M is a positive number.

[0143] Further, such as Figure 4 As shown, the execution unit 304 includes:

[0144] A calling module 3041 is used to call the target operator to process the current production data;

[0145] An output module 3042 is used to output the prediction results corresponding to each target operator;

[0146] A composition module 3043 is used to combine the prediction results corresponding to each of the target operators into a prediction result set;

[0147] Feedback module 3044 is used to feed back the prediction result set to the user.

[0148] Further, such as Figure 4 As shown, the device also includes:

[0149] The parsing unit 311 is used for parsing the parameter dimensions included in the historical production data before screening the multiple operators using the historical production data to select the operator matching the target oil field from the multiple operators;

[0150] A judging unit 312, configured to compare the parameter dimension obtained by the parsing unit 311 with the parameter dimension to be processed by the operator to judge whether they match;

[0151] The discarding unit 313 is used to discard the operator when the judging unit 312 judges that the parameter dimension obtained by the parsing unit 311 matches the parameter dimension that the operator needs to process.

[0152] Further, such as Figure 4 As shown, the device also includes:

[0153] A prediction unit 314 is used for, before screening the plurality of operators using the historical production data and selecting an operator matching the target oil field from the plurality of operators, outputting a prediction result of each operator on the future production of the target oil field by adapting the historical production data to each operator;

[0154] The search unit 308 is further used to search for the historical actual result corresponding to the target oil field in the same time interval according to the time interval corresponding to the prediction result;

[0155] The judging unit 312 is further configured to judge whether a prediction error reaches a preset threshold by comparing the prediction result predicted by each operator with the historical actual result found by the finding unit 308;

[0156] The discarding unit 313 is further configured to discard the operator when the judging unit 312 judges that the prediction error reaches a preset threshold.

[0157] In summary, the embodiments of the present invention provide a method and device for predicting oil field production. The embodiments of the present invention use a machine learning algorithm to automatically find a better model suitable for the target oil field production prediction from a large number of known oil field production prediction models. Compared with the prior art, it solves the problem of manually finding a prediction method and model suitable for the target oil field from a large number of known prediction methods, which leads to high difficulty and workload, complex calculations, and practical difficulties. The method provided by the embodiments of the present invention, for each target oil field to be tested, when receiving the current production data of the target oil field uploaded by the user, automatically and efficiently finds the better model for executing the prediction of the oil field production, and calls this better model to predict the future production of the target oil field, and also ensures the accuracy of the predicted oil field production.

[0158] The oilfield production prediction device includes a processor and a memory. The receiving unit, acquisition unit, screening unit and execution unit are all stored in the memory as program units. The processor executes the program units stored in the memory to realize corresponding functions.

[0159] The processor includes a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and the machine learning algorithm can be used to automatically find the best model suitable for the target oil field production prediction from many known oil field production prediction models by adjusting the kernel parameters, and also ensure the accuracy of the predicted oil field production.

[0160] An embodiment of the present invention provides a storage medium on which a program is stored. When the program is executed by a processor, the method for predicting oil field production is implemented.

[0161] An embodiment of the present invention provides a processor, which is used to run a program, wherein the method for predicting oil field production is executed when the program is run.

[0162] An embodiment of the present invention provides an electronic device 40, such as Figure 5 As shown, the device includes at least one processor 401, and at least one memory 402 and a bus 403 connected to the processor 401; wherein the processor 401 and the memory 402 communicate with each other via the bus 403; the processor 401 is used to call the program instructions in the memory 402 to execute the above-mentioned oil field production prediction method.

[0163] The devices in this article can be servers, PCs, PADs, mobile phones, etc.

[0164] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program for initializing the following method steps:

[0165] A method for predicting oil field production, the method comprising: receiving an oil field production prediction instruction, the prediction instruction carrying a target oil field identifier and current production data of the target oil field; obtaining historical production data corresponding to the target oil field according to the target oil field identifier; obtaining multiple oil field production prediction models, each of which is encapsulated into an independent operator; screening multiple operators using the historical production data, and screening out an operator matching the target oil field from the multiple operators; processing the current production data by calling the operator matching the target oil field, executing an operation of predicting the future production of the target oil field, and outputting a prediction result.

[0166] Furthermore, the method of using the historical production data to screen the multiple operators and selecting an operator matching the target oil field from the multiple operators includes: using the historical production data to train the multiple operators respectively to obtain the optimal parameters corresponding to each operator; and adjusting the corresponding operator according to the optimal parameters to obtain the operator matching the production of the target oil field.

[0167] Furthermore, after selecting an operator matching the target oil field from multiple operators, if there are multiple operators matching the target oil field, the method further includes: if there are multiple operators matching the target oil field, determining the operator matching the target oil field as a candidate operator; dividing the historical production data into partial production data corresponding to each preset time unit according to preset time units; using each candidate operator to process the partial production data to obtain a prediction result corresponding to each time unit; searching for the historical actual result corresponding to each time unit; calculating the average error percentage obtained by each candidate operator processing the historical production data based on the prediction result and the historical actual result corresponding to each time unit; selecting the operator whose average error percentage meets the preset conditions as the target operator, the preset conditions being the top N operators with the smallest average error percentage, or the top M% operators with the smallest average error percentage, where N is a positive integer and M is a positive number.

[0168] Furthermore, the method processes the current production data by calling the operator matching the target oil field, executes the operation of predicting the future production of the target oil field, and outputs the prediction result, including: calling the target operator to process the current production data, and outputting the prediction result corresponding to each of the target operators; forming the prediction results corresponding to each of the target operators into a prediction result set; and feeding back the prediction result set to the user.

[0169] Furthermore, before using the historical production data to screen the multiple operators and selecting the operator matching the target oil field from the multiple operators, the method also includes: parsing the parameter dimensions contained in the historical production data; comparing the parameter dimensions with the parameter dimensions that the operator needs to process to determine whether they match; if not, discarding the operator.

[0170] Furthermore, before using the historical production data to screen the multiple operators and selecting the operator matching the target oil field from the multiple operators, the method also includes: outputting the prediction result of each operator on the future production of the target oil field by adapting the historical production data to each operator; obtaining the historical actual results corresponding to the target oil field in the same time interval according to the time interval corresponding to the prediction result; judging whether the prediction error reaches a preset threshold by comparing the prediction result calculated by each operator with the historical actual result; if so, discarding the operator.

[0171] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes 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 generate 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.

[0172] In a typical configuration, the device includes one or more processors (CPU), memory and bus. The device may also include input / output interface, network interface and the like.

[0173] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip. The memory is an example of a computer-readable medium.

[0174] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0175] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0176] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt 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.) that contain computer-usable program code.

[0177] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for predicting oil field production, characterized in that: The method comprises: receiving an oilfield production prediction instruction, wherein the prediction instruction carries a target oilfield identifier and current production data of the target oilfield; According to the target oil field identifier, obtaining the historical production data corresponding to the target oil field; Acquire multiple oilfield production prediction models, each of which is encapsulated into an independent operator; The historical production data is used to screen the multiple operators, and the operator matching the target oil field is screened out from the multiple operators, including: using the historical production data to train the multiple operators respectively to obtain the optimal parameters corresponding to each operator; adjusting the corresponding operator according to the optimal parameters to obtain the operator matching the production of the target oil field; If there are multiple operators matching the target oil field, the operators matching the target oil field are determined as operators to be selected; According to a preset time unit, the historical production data is divided into partial production data corresponding to each preset time unit; Using each of the operators to be selected to process the part of the production data, to obtain a prediction result corresponding to each of the time units; Find the historical actual results corresponding to each of the time units; Calculate the average error percentage of each of the to-be-selected operators in processing the historical production data according to the prediction results and historical actual results corresponding to each of the time units; An operator whose average error percentage meets a preset condition is selected as a target operator, wherein the preset condition is the first N operators with the smallest average error percentage, or the first M% operators with the smallest average error percentage, where N is a positive integer and M is a positive number; The target operator matching the target oil field is called to process the current production data, an operation of predicting the future production of the target oil field is performed, and a prediction result is output, including: calling the target operator to process the current production data, outputting the prediction result corresponding to each target operator; forming a prediction result set from the prediction results corresponding to each target operator; and feeding back the prediction result set to a user.

2. The method according to claim 1, characterized in that: Before screening the plurality of operators by using the historical production data to select an operator matching the target oil field from the plurality of operators, the method further includes: Analyzing the parameter dimensions included in the historical production data; By comparing the parameter dimension with the parameter dimension that the operator needs to process, determining whether they match; If there is no match, the operator is discarded.

3. The method according to claim 1, characterized in that Before screening the plurality of operators by using the historical production data to select the target operator matching the target oil field from the plurality of operators, the method further includes: By adapting the historical production data to each of the operators, outputting the prediction result of each of the operators on the future production of the target oil field; According to the time interval corresponding to the prediction result, obtaining the historical actual result corresponding to the target oil field in the same time interval; By comparing the prediction result calculated by each operator with the historical actual result, it is determined whether the prediction error reaches a preset threshold; If so, the operator is discarded.

4. A device for predicting oil field production, characterized in that: The device comprises: A receiving unit, used for receiving an oilfield production prediction instruction, wherein the prediction instruction carries a target oilfield identifier and current production data of the target oilfield; an acquisition unit, configured to search for historical production data corresponding to the target oil field according to the target oil field identifier received by the receiving unit; The acquisition unit is further used to acquire multiple oilfield production prediction models, each of which is encapsulated into an independent operator; a screening unit, configured to screen the operators acquired by the acquisition unit using the historical production data received by the receiving unit, and select an operator matching the target oil field from the multiple operators; The screening unit includes: a training module, which is used to train multiple operators respectively using historical production data to obtain the optimal parameters corresponding to each operator; a determination module, which is used to adjust the corresponding operator according to the optimal parameters obtained by the training module to determine the operator that matches the production of the target oil field; The determination unit is further configured to, after selecting an operator matching the target oil field from the plurality of operators, determine the operator matching the target oil field as the operator to be selected if there are multiple operators matching the target oil field; A division unit, used for dividing the historical production data into partial production data corresponding to each preset time unit according to the preset time unit; A processing unit, used to process the part of the production data using each of the operators to be selected to obtain a prediction result corresponding to each of the time units; A search unit, used to obtain the historical actual result corresponding to each of the time units; A calculation unit, configured to calculate an average error percentage obtained by each of the to-be-selected operators in processing the historical production data according to the prediction result obtained by the processing unit corresponding to each of the time units and the historical actual result obtained by the obtaining unit; A selection unit, used for selecting an operator whose average error percentage calculated by the calculation unit meets a preset condition as a target operator, wherein the preset condition is the first N operators with the smallest average error percentage, or the first M% operators with the smallest average error percentage, wherein N is a positive integer and M is a positive number; An execution unit, configured to process the current production data by calling the operator matched with the target oil field screened by the screening unit, execute an operation of predicting the future production of the target oil field, and output a prediction result; Wherein, the execution unit includes: A calling module, used for calling the target operator to process the current production data; An output module, used to output the prediction results corresponding to each target operator; A composition module, used for composing the prediction results corresponding to each of the target operators into a prediction result set; The feedback module is used to feed back the prediction result set to the user.

5. A storage medium, characterized in that: The storage medium includes a stored program, wherein when the program is executed, the device where the storage medium is located is controlled to execute the method for predicting oil field production according to any one of claims 1 to 3.

6. An electronic device, characterized in that: The device includes at least one processor, and at least one memory and a bus connected to the processor; Wherein, the processor and the memory communicate with each other through the bus; The processor is used to call the program instructions in the memory to execute the oil field production prediction method according to any one of claims 1-3.

Images