CN113010501A

CN113010501A - Recovery ratio prediction model obtaining method, recovery ratio prediction method and product

Info

Publication number: CN113010501A
Application number: CN201911317079.XA
Authority: CN
Inventors: 吴文旷
Original assignee: Beijing Gridsum Technology Co Ltd
Current assignee: Beijing Gridsum Technology Co Ltd
Priority date: 2019-12-19
Filing date: 2019-12-19
Publication date: 2021-06-22

Abstract

The application relates to an oil reservoir recovery ratio prediction model obtaining method, an oil reservoir recovery ratio prediction method and a product. The method comprises the following steps: acquiring training characteristic parameters and testing characteristic parameters; the training characteristic parameters and the testing characteristic parameters comprise static parameters representing fixed characteristics of the oil deposit and dynamic parameters representing variation characteristics of the oil deposit; training a plurality of submodels to be trained respectively by adopting training characteristic parameters to obtain a plurality of submodels to be verified; the sub-models to be verified are models based on different algorithms; respectively inputting the test characteristic parameters into each submodel to be verified to obtain a test result of each submodel to be verified; and screening a plurality of target submodels from the plurality of submodels to be verified according to the test result of each submodel to be verified and the preset screening requirements to obtain an oil reservoir recovery ratio prediction model. The method can improve the accuracy of the prediction result of the recovery ratio.

Description

Recovery ratio prediction model obtaining method, recovery ratio prediction method and product

Technical Field

The application relates to the technical field of oil exploitation, in particular to a recovery ratio prediction model obtaining method, a recovery ratio prediction method and a product.

Background

In the process of oil exploitation, the recovery ratio is an important parameter for evaluating the development effect of an oil field, and the method accurately predicts the recovery ratio of crude oil, and is the basis for scientifically managing the oil field, compiling an economically feasible development scheme and making the production capacity scale. If the recovery rate can be accurately predicted, old oil fields which are put into production and new oil fields which are not put into production can be better managed, the oil extraction process is reasonably controlled and optimized, more oil is extracted from underground, and the development effect of the oil fields is improved.

At present, oil recovery prediction methods for oil reservoirs or oil fields are various and comprise a static method and a dynamic method, wherein the static method is mainly used for determining the oil recovery rate by adopting static methods such as a theoretical formula, an empirical formula, an analogy and the like aiming at the oil field which is not developed and put into production or the oil field at the initial development stage; the dynamic method is the most common method in the calculation of the recoverable reserves. The dynamic method is mainly used for predicting the dynamic trend of future development according to the mining historical dynamic data and the change rule of an oil reservoir, and can be divided into a decreasing curve method, a water drive characteristic curve method, a child constitution diagram version method, an injection and production relation method, a numerical simulation method and the like. Conventional methods of predicting recovery are each applicable to different reservoir types.

Therefore, the traditional recovery factor prediction method has limitation, so that the prediction result of the recovery factor is inaccurate.

Disclosure of Invention

In view of the above, it is necessary to provide a method for obtaining a reservoir recovery prediction model, a method for predicting reservoir recovery, an apparatus, a computer device, and a storage medium, which can improve accuracy.

In a first aspect, an embodiment of the present application provides a method for obtaining a reservoir recovery prediction model, the method including:

acquiring training characteristic parameters and testing characteristic parameters; the training characteristic parameters and the testing characteristic parameters comprise static parameters representing fixed characteristics of the oil deposit and dynamic parameters representing variation characteristics of the oil deposit;

respectively training a plurality of submodels to be trained by adopting the training characteristic parameters to obtain a plurality of submodels to be verified; the sub-models to be verified are models based on different algorithms, and each sub-model to be trained is a sub-model matched with the type of the training characteristic parameter;

respectively inputting the test characteristic parameters into each sub-model to be verified to obtain a test result of each sub-model to be verified;

screening a plurality of target sub-models from the plurality of sub-models to be verified according to the test result of each sub-model to be verified and preset screening requirements to obtain the plurality of target sub-models serving as the oil reservoir recovery rate prediction model; the oil reservoir recovery rate prediction model is used for outputting oil reservoir recovery rate prediction results according to input characteristic parameters to be predicted, and the oil reservoir recovery rate prediction results are determined according to the recovery rate predicted by each target sub-model.

In a second aspect, embodiments of the present application provide a method for reservoir recovery prediction, the method comprising:

acquiring a characteristic parameter to be predicted; the characteristic parameters to be predicted comprise static parameters representing fixed characteristics of the oil reservoir and dynamic parameters representing variation characteristics of the oil reservoir;

processing the characteristic parameters to be predicted by adopting the oil reservoir recovery model in any one of the embodiments to output an oil reservoir recovery prediction result;

the oil reservoir recovery model comprises a plurality of target sub-models, and the oil reservoir recovery prediction result is determined according to the recovery predicted by each target sub-model.

In one embodiment, the data format of the predicted recovery factor of each target sub-model is a single value or a numerical range, and the data format of the predicted reservoir recovery factor is a numerical range;

in one embodiment, when the data format of the predicted recovery factor of each target sub-model is a single value, the reservoir recovery factor prediction result is determined according to the predicted recovery factor of each target sub-model, and the method comprises the following steps:

respectively determining a minimum value and a maximum value from the single values of the recovery factor predicted by each target sub-model, and determining a numerical range between the minimum value and the maximum value as a prediction result of the recovery factor of the oil reservoir;

in one embodiment, when the data format of the predicted recovery factor of each target sub-model is a numerical range, the reservoir recovery factor prediction result is determined according to the predicted recovery factor of each target sub-model, and the method comprises the following steps:

respectively determining a minimum value and a maximum value from the numerical range of the recovery factor predicted by each target sub-model, and determining the numerical range between the minimum value and the maximum value as the prediction result of the recovery factor of the oil reservoir;

and taking the average value of the single recovery factor value predicted by each target sub-model as the central value of the prediction range of the recovery factor of the oil reservoir, taking the mean square deviation of each single recovery factor value and the central value as the floating value of the prediction range of the recovery factor of the oil reservoir, and determining the numerical range obtained by increasing and decreasing the floating value of the central value as the prediction result of the recovery factor of the oil reservoir.

In a third aspect, embodiments of the present application provide an apparatus for obtaining a reservoir recovery prediction model, the apparatus including:

the acquisition module is used for acquiring training characteristic parameters and testing characteristic parameters; the training characteristic parameters and the testing characteristic parameters comprise static parameters representing fixed characteristics of the oil deposit and dynamic parameters representing variation characteristics of the oil deposit;

the training module is used for respectively training a plurality of submodels to be trained by adopting the training characteristic parameters to obtain a plurality of submodels to be verified; the sub-models to be verified are models based on different algorithms, and each sub-model to be trained is a sub-model matched with the type of the training characteristic parameter;

the test module is used for respectively inputting the test characteristic parameters into each submodel to be verified to obtain a test result of each submodel to be verified;

the processing module is used for screening a plurality of target sub-models from the plurality of sub-models to be verified according to the test result of each sub-model to be verified and preset screening requirements to obtain the plurality of target sub-models as the oil reservoir recovery rate prediction models;

the oil reservoir recovery rate prediction model is used for outputting oil reservoir recovery rate prediction results according to input characteristic parameters to be predicted, and the oil reservoir recovery rate prediction results are determined according to the recovery rate predicted by each target sub-model.

In a fourth aspect, embodiments of the present application provide a reservoir recovery prediction device, the device comprising:

the acquisition module is used for acquiring the characteristic parameters to be predicted; the characteristic parameters to be predicted comprise static parameters representing fixed characteristics of the oil reservoir and dynamic parameters representing variation characteristics of the oil reservoir;

the prediction module is used for processing the characteristic parameters to be predicted in the oil reservoir recovery model according to any one of the embodiments and outputting a oil reservoir recovery prediction result; the oil reservoir recovery model comprises a plurality of target sub-models, and the oil reservoir recovery prediction result is determined according to the recovery predicted by each target sub-model.

In a fifth aspect, an embodiment of the present application provides a computer device, including a memory and a processor, where the memory stores a computer program, and the processor implements the following steps when executing the computer program:

In a sixth aspect, an embodiment of the present application provides a computer device, including a memory and a processor, where the memory stores a computer program, and the processor implements the following steps when executing the computer program:

processing the characteristic parameters to be predicted by adopting the oil reservoir recovery model in any embodiment to output an oil reservoir recovery prediction result;

In a seventh aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the following steps:

In an eighth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the following steps:

The recovery factor prediction model obtaining method, the recovery factor prediction device, the computer equipment and the storage medium are characterized in that training characteristic parameters and testing characteristic parameters are obtained, the training characteristic parameters are adopted to respectively train a plurality of submodels to be trained to obtain a plurality of submodels to be verified, the testing characteristic parameters are respectively input into each submodel to be verified to obtain a testing result of each submodel to be verified, finally, according to the testing result of each submodel to be verified, a plurality of target submodels are obtained by screening from the plurality of submodels to be verified according to preset screening requirements, the oil reservoir recovery factor prediction model is used for outputting an oil reservoir recovery factor prediction result according to the input characteristic parameters to be predicted, the oil reservoir recovery factor prediction result is determined according to the recovery factor predicted by each target submodel, and therefore the oil reservoir recovery factor prediction model consisting of a plurality of target submodels capable of obtaining the optimal testing result can be obtained And (4) modeling. Because the plurality of target sub-models included in the oil reservoir recovery ratio prediction model are the target sub-models which are obtained through screening and have the optimal test results, namely, the target sub-models with higher precision, the technical problem that the application scene is single due to the fact that a single model is adopted for recovery ratio prediction in the traditional method can be solved, the application scene of the model is greatly enriched, and the robustness of the model is improved. Meanwhile, the problems of large calculated amount and complex calculation caused by manual selection of a prediction model are avoided, a large amount of time and labor cost are saved, the prediction efficiency is greatly improved, the prediction precision of the oil recovery ratio is improved, and the use threshold of the recovery ratio prediction method is lowered. Optionally, the multiple sub-models to be verified may include existing business mechanism models, such as a water drive characteristic curve, an injection-production relationship model, and the like, and may further include models based on machine learning and deep learning, and by combining these models together, a comprehensive model combining business and artificial intelligence is formed, so that the obtained oil reservoir recovery prediction model is more reasonable.

Drawings

FIG. 1 is a diagram illustrating an internal structure of a computer device according to an embodiment;

FIG. 2 is a schematic flow diagram of a method for obtaining a reservoir recovery prediction model according to one embodiment;

FIG. 3 is a schematic flow diagram of a method for obtaining a reservoir recovery prediction model according to another embodiment;

FIG. 4 is a schematic flow diagram of a method for reservoir recovery prediction model acquisition according to yet another embodiment;

FIG. 5 is a schematic flow diagram of a reservoir recovery prediction method provided by an embodiment;

FIG. 6 is a schematic diagram of an embodiment of a reservoir recovery prediction model acquisition device;

fig. 7 is a schematic diagram of a reservoir recovery prediction device according to an embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

The method for acquiring the oil reservoir recovery ratio prediction model and the oil reservoir recovery ratio prediction method provided by the embodiment of the application can be applied to the computer equipment shown in fig. 1. The computer device comprises a processor, a memory, a network interface, a database, a display screen and an input device which are connected through a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used to store the reservoir recovery prediction model in the following examples, and the specific description of the reservoir recovery prediction model is provided in the following examples. The network interface of the computer device may be used to communicate with other devices outside over a network connection. Optionally, the computer device may be a server, a desktop, a personal digital assistant, other terminal devices such as a tablet computer, a mobile phone, and the like, or a cloud or a remote server, and the specific form of the computer device is not limited in the embodiment of the present application. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like. Of course, the input device and the display screen may not belong to a part of the computer device, and may be external devices of the computer device.

To facilitate understanding, the terms referred to herein are explained and illustrated:

recovery ratio: refers to the ratio of the amount of crude oil produced to the original geological reserve of the reservoir, usually expressed as a percentage. The oil field development practical data shows that the reservoir recovery factor is closely related to natural conditions such as reservoir energy type and size, reservoir lithology change and heterogeneous condition, formation crude oil physical property and the like, and also has great relation to technical measures and production management methods for oil field development and exploitation.

Automatic machine learning (AutoML): the application of machine learning requires a large amount of manual intervention, which is manifested in: and (3) various aspects of machine learning such as feature extraction, model selection and parameter adjustment. Automated machine learning attempts to automatically learn these important steps related to features, models, optimizations, evaluations, so that the machine learning model can be applied without human intervention.

Oil field: it is a general term for a specific region of crude oil production, sometimes referred to as an underground accumulated oil layer in a specific region. Several oil zones may be broadly referred to together as an oil field. Such as Daqing oil field, North sea oil field in England, autumn Bright oil field in Russia, etc. Oil fields are naturally occurring hydrocarbons in the subsurface that are liquid at surface conditions. In contrast, it is still a gas at surface conditions, and is natural gas. A particular area of natural gas production is a natural gas field. The size of the petroleum recoverable reserve determines the value of exploitation, and therefore accurate calculation of the size of the oil-bearing area, the number and thickness of oil layers, the petroleum reserve per unit area and the like is required. The oil field with the reserves of more than 5 hundred million tons is a super-huge oil field, the oil field with the reserves of more than 7000 million tons to more than 1 hundred million tons is a large oil field, and the oil field with the reserves of less than 7000 million tons is a medium-small oil field. The possible oil field and annual output are calculated, and some oil fields have large reserves but not necessarily high output and are mainly influenced by the driving capability of the oil fields. Normally, the reserves to the production from the field need to be carefully calculated, indeed economically valuable, to be formally exploited.

Oil field development: on the basis of oil field geology and dynamic research, oil gas resources in a stratum are extracted from the underground to the ground through development, deployment, implementation and other engineering technical measures so that the oil gas resources can be used as the whole process of commodity gas output.

Adjusting parameters: parameters in the model (such as the number of layers of the neural network in the deep learning model) are adjusted and optimized. The final goal of tuning is to make the model predicted yield after training more accurate, closer to one step in the direction of least error, i.e., to make the loss function as small as possible.

Those skilled in the art will appreciate that the architecture shown in fig. 1 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

The following describes the technical solutions of the present application and how to solve the above technical problems with specific examples. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.

It should be noted that the execution subject of the method embodiments described below may be a reservoir recovery prediction model obtaining device or a reservoir recovery prediction device, which may be implemented as part of or all of the computer device described above by software, hardware or a combination of software and hardware. The following method embodiments are described by taking the execution subject as the computer device as an example.

Fig. 2 is a schematic flow chart of a reservoir recovery prediction model obtaining method according to an embodiment. The embodiment relates to a specific process for automatically training computer equipment to obtain a reservoir recovery ratio prediction model. As shown in fig. 2, includes:

s10, acquiring training characteristic parameters and testing characteristic parameters; the training characteristic parameters and the testing characteristic parameters respectively comprise static parameters representing fixed characteristics of the oil deposit and dynamic parameters representing variation characteristics of the oil deposit.

Specifically, the computer device may read training feature parameters and testing feature parameters stored in the device, may also receive training feature parameters and testing feature parameters input by other devices or a user, and may also perform preprocessing on the sample feature parameters to obtain the training feature parameters and the testing feature parameters, where optionally, the preprocessing includes, but is not limited to, performing processing such as reasonableness screening and interpolation on data. The training characteristic parameters are characteristic parameters used for model training, and the testing characteristic parameters are used for testing whether the output result of the trained model meets the precision requirement. It should be noted that the number of the training characteristic parameters and the number of the testing characteristic parameters are multiple, the greater the number of the samples of the training characteristic parameters, the stronger the robustness of the trained model, and the greater the number of the samples of the testing characteristic parameters, the closer the output result of the model is to the reality. The training characteristic parameters and the testing characteristic parameters comprise static parameters and dynamic parameters, wherein the static parameters are parameters for representing fixed characteristics of an oil reservoir, and include but are not limited to oil-water viscosity ratio, crude oil volume coefficient, crude oil viscosity, formation water viscosity, average air permeability, effective permeability, average effective thickness of an oil layer, well control area, permeability variation coefficient, oil layer temperature, effective porosity, well pattern density, main lithology of the oil reservoir, bottom water energy, injection-production well ratio, oil layer communication rate and the like; the dynamic parameters include, but are not limited to, the year and month of production, the annual oil production, the annual water cut, the annual water injection, the cumulative oil production, the cumulative water production, and other data of the oil reservoir. Optionally, the dynamic data and the static data can be made into a data acquisition template, the number of columns in the template is fixed, the number of rows can be added by a user on line or off line according to the actual oil reservoir condition, and the data columns without values can be emptied, so that the processed dynamic data and the processed static data are convenient to process.

S20, training the sub-models to be trained respectively by adopting the training characteristic parameters to obtain a plurality of sub-models to be verified; the sub-models to be verified are models based on different algorithms; each submodel to be trained is a submodel matched with the type of the training characteristic parameter;

specifically, the computer device inputs the training characteristic parameters into a plurality of submodels to be trained, trains each submodel to be trained separately, and automatically adjusts and optimizes parameters through each submodel to be trained to obtain a training result. In the training process of each submodel to be trained, the computer equipment calculates the parameters or the hyper-parameters corresponding to the model with the minimum loss function according to the training result, so that each submodel to be trained is updated, a plurality of trained submodels to be verified are obtained, and each submodel to be trained can be trained to obtain one trained submodel to be verified. It should be noted that the submodel to be trained is a model based on different algorithms, and may include, but is not limited to: at least two of a static model, a machine learning model, a deep learning algorithm model, a deep convolutional neural network model, a cyclic neural network model and a long-short term memory network model, so that the plurality of to-be-verified submodels obtained by training are also models based on different algorithms. Meanwhile, each of the submodels to be trained is a submodel matched with the type of the training characteristic parameter, that is, different types of training characteristic parameters correspond to different submodels to be trained, for example, when the training characteristic parameters are static data, the corresponding types of the submodels to be trained are static models.

And S30, respectively inputting the test characteristic parameters into each sub-model to be verified to obtain the test result of each sub-model to be verified.

Specifically, the computer device inputs the test characteristic parameters into each sub-model to be verified, and each sub-model to be verified outputs test results of the test characteristic parameters, and the test results can represent the degree of the accuracy of the test characteristic parameters of each sub-model to be verified. For example, for a set of test characteristic parameters, the test result of the sub-model to be verified may be a value of a loss function, and when the value of the loss function is large, the test result characterizing the sub-model to be verified is superior, that is, the sub-model to be verified is more accurate; when the value of the loss function is small, the test result for characterizing the sub-model to be verified is poor, that is, the accuracy of the sub-model to be verified is low.

S40, according to the test result of each sub-model to be verified, according to the preset screening requirement, screening a plurality of target sub-models from the plurality of sub-models to be verified to obtain a plurality of target sub-models serving as the oil reservoir recovery rate prediction model; the oil reservoir recovery rate prediction model is used for outputting oil reservoir recovery rate prediction results according to input characteristic parameters to be predicted, and the oil reservoir recovery rate prediction results are determined according to the recovery rate predicted by each target sub-model.

Specifically, the computer device screens a plurality of target sub-models from the plurality of sub-models to be verified according to the test result of each sub-model to be verified and preset screening requirements, and uses the plurality of target sub-models as the oil reservoir recovery ratio prediction model. It should be noted that the screening requirement may be that the test result satisfies a preset threshold, or the test results are arranged in the first few names, or the submodels to be verified corresponding to the test results are used as the target submodels by selecting the submodels to be ranked in the first few names from the test results according to a certain proportion. For example, the preset number of the loss function values in the test result, for example, three to-be-verified sub-models, may be screened to obtain a first target sub-model, a second target sub-model, and a third target sub-model, and then the first target sub-model, the second target sub-model, and the third target sub-model are combined to serve as the reservoir recovery factor prediction model.

In this embodiment, the computer device obtains the training characteristic parameters and the testing characteristic parameters, and trains the plurality of submodels to be trained respectively by using the training characteristic parameters to obtain a plurality of submodels to be verified, then the test characteristic parameters are respectively input into each submodel to be verified to obtain the test result of each submodel to be verified, and finally according to the test result of each submodel to be verified, screening a plurality of target submodels from a plurality of submodels to be verified according to preset screening requirements, because the oil reservoir recovery rate prediction model is used for outputting the oil reservoir recovery rate prediction result according to the input characteristic parameters to be predicted, the oil reservoir recovery rate prediction result is determined according to the recovery rate predicted by each target sub-model, therefore, a reservoir recovery prediction model consisting of a plurality of target sub-models which can obtain the optimal test result can be obtained. Because the plurality of target sub-models included in the oil reservoir recovery ratio prediction model are the target sub-models which are obtained through screening and have the optimal test results, namely, the target sub-models with higher precision, the technical problem that the application scene is single due to the fact that a single model is adopted for recovery ratio prediction in the traditional method can be solved, the application scene of the model is greatly enriched, and the robustness of the model is improved. Meanwhile, the problems of large calculated amount and complex calculation caused by manual selection of a prediction model are avoided, a large amount of time and labor cost are saved, the prediction efficiency is greatly improved, the prediction precision of the oil recovery ratio is improved, and the use threshold of the recovery ratio prediction method is lowered. Optionally, the multiple sub-models to be verified may include existing business mechanism models, such as a water drive characteristic curve, an injection-production relationship model, and the like, and may further include models based on machine learning and deep learning, and by combining these models together, a comprehensive model combining business and artificial intelligence is formed, so that the obtained oil reservoir recovery prediction model is more reasonable.

Optionally, the computer device may issue the obtained multiple target sub-models into an API in a micro-service manner for a service user to directly call, so that the service user may upload characteristic parameters of the user who needs to perform prediction, and directly call the optimal prediction model to perform recovery factor prediction, which is more convenient and faster to use.

Optionally, on the basis of the foregoing embodiment, a possible implementation manner of the step S10 may be as shown in fig. 3, and includes:

s11, obtaining initial characteristic parameters; wherein the initial characteristic parameters comprise static parameters and dynamic parameters.

And S12, carrying out data cleaning on the initial characteristic parameters to obtain intermediate characteristic parameters.

Specifically, the computer device obtains the initial characteristic parameter, for example, the initial characteristic parameter may be a static parameter and a dynamic parameter automatically imported by the device, for example, the initial characteristic parameter is downloaded from a database, or the initial characteristic parameter input by the computer device is manually input by an engineer, which is not limited in this embodiment. The computer device performs data cleaning on the received initial characteristic parameters, may be, automatically verifies the collected initial characteristic parameters, and makes a data cleaning program according to actual business rules for ranges of different data column porosities, and may include clearing abnormal points in the initial characteristic parameters, such as parameters with porosity greater than 100% and water content less than 0%, deleting, smoothing or correcting inconsistent or contradictory data points in the initial characteristic parameters, interpolating data points that can pass data interpolation, and the like, thereby obtaining intermediate characteristic parameters.

S13, splitting the intermediate characteristic parameters to obtain the training characteristic parameters and the testing characteristic parameters; the training characteristic parameters and the testing characteristic parameters comprise static parameters and dynamic parameters.

Specifically, the computer device splits the intermediate characteristic parameters, and uses a part of the intermediate characteristic parameters as training characteristic parameters and a part of the intermediate characteristic parameters as testing characteristic parameters. It should be noted that, in the process of splitting the intermediate characteristic parameter, the dynamic data in the intermediate characteristic parameter may be split, and the static data in the intermediate characteristic parameter may be split, so that the obtained training characteristic parameter and the test characteristic parameter both include a dynamic parameter and a dynamic parameter.

Optionally, one possible implementation manner of this step S13 may include: splitting the intermediate characteristic parameters according to a preset proportional relation to obtain the training characteristic parameters and the testing characteristic parameters which are matched with the proportional relation. Alternatively, the proportional relationship may be that the data amount of the training characteristic parameter and the test characteristic parameter is 80% and 20%, respectively, and may also be other ratios, such as five ratios and the like, and the proportional relationship may be set as required. The intermediate characteristic parameters are split according to the preset proportional relation to obtain the training characteristic parameters and the testing characteristic parameters matched with the proportional relation, so that the training characteristic parameters and the testing characteristic parameters matched with the proportional relation are obtained, the training data and the testing data are reasonably distributed, and the obtained oil reservoir recovery ratio prediction model is more accurate.

Optionally, before this step S13, the method may further include: judging whether the number of the initial characteristic parameters is smaller than a preset parameter number threshold value or not; if yes, the proportional relation between the training characteristic parameters and the testing characteristic parameters is 1 to 0. For example, when the number of the initial feature parameters is less than the preset parameter number threshold 5, the splitting operation is not performed, that is, the initial feature parameters are subjected to data cleaning and are all classified as training feature parameters, that is, the proportional relationship between the training feature parameters and the test feature parameters is set to 1 to 0. Optionally, at this time, the computer device may download other test feature parameters from the platform, or may select a part from the training feature parameters as the test feature parameters, which is not limited in this embodiment. Optionally, the parameter number threshold may also be 8, 10 or other data. In this embodiment, when the number of the initial feature parameters is smaller than the preset parameter number threshold, the computer device sets the ratio of the training feature parameters to the testing feature parameters to 1 to 0, so as to ensure the number of samples of the training feature parameters, thereby improving the accuracy of the trained model.

In the embodiment shown in fig. 3, the computer device obtains the intermediate characteristic parameters by obtaining the initial characteristic parameters and performing data cleaning on the initial characteristic parameters, so that unreasonable data is deleted, the validity of the data is improved, and the accuracy of the trained model is further improved; meanwhile, the computer equipment splits the intermediate characteristic parameters to obtain training characteristic parameters and testing characteristic parameters, so that the training characteristic parameters and the testing characteristic parameters can be respectively used for training and testing the model, and the accuracy of the oil reservoir recovery ratio prediction model is further improved.

Optionally, on the basis of the foregoing embodiments, before the step S20, as shown in fig. 4, the method further includes:

s21, obtaining a plurality of initial prediction models; the initial prediction submodel is a model for predicting the recovery ratio of the oil reservoir based on the characteristic parameters of different data types.

And S22, screening out a plurality of submodels to be trained from the plurality of initial prediction submodels according to the data types of the training characteristic parameters and the testing characteristic parameters.

Specifically, the computer device obtains a plurality of initial prediction models, each of which is capable of predicting recovery from a reservoir under different conditions, and the plurality of initial prediction models include models of a plurality of algorithms, each of which may match a different type of characteristic parameter. And training characteristic parameters and the data types of the test characteristic parameters by the computer equipment, automatically filtering part of initial prediction models with obviously improper recovery factor prediction from the training characteristic parameters and the data types of the test characteristic parameters by a Bayesian network optimization method, screening part of initial prediction models matched with the training characteristic parameters and the data types of the test characteristic parameters from a plurality of initial prediction models, and using the part of initial prediction models as the sub-models to be trained. Therefore, part of the initial prediction models which are obviously unsuitable for recovery factor prediction can be automatically filtered, so that invalid calculation is avoided, the model training efficiency is greatly improved, and the efficiency is also greatly improved in the subsequent recovery factor prediction.

Optionally, on the basis of the foregoing embodiment, one possible implementation manner of step S40 may further include: sequencing each test result according to the error rate; and screening a preset number of submodels to be verified from the error rate sequence to be used as the target submodels. Specifically, the computer device may sort all the test results according to the error rate, and screen out a preset number of sub-models to be verified as the target sub-models, for example, the sub-models to be verified may be sorted according to the error rate from low to high, and the top three sub-models to be verified are used as the target sub-models, so that a good balance between accuracy and robustness can be achieved. In this embodiment, the computer device sorts each of the test results according to the error rate; and screening a preset number of sub-models to be verified from the error rate sequence to be used as the target sub-models, so that the most accurate sub-models to be verified are screened to be used as the target sub-models, the accuracy of the models is greatly improved, and the prediction accuracy of the oil reservoir recovery rate can be improved.

Alternatively, the data format of the predicted recovery factor of each target sub-model may be a single value, and the data format of the reservoir recovery factor prediction result may be a range of values. Determining the oil reservoir recovery rate prediction result according to the recovery rate predicted by each target sub-model, wherein the method specifically comprises the following steps:

and respectively determining a minimum value and a maximum value from the single values of the recovery factor predicted by each target sub-model, and determining a numerical range between the minimum value and the maximum value as a prediction result of the recovery factor of the oil reservoir. For example, the predicted recovery single value of the target sub-model includes A, B and C, wherein A is maximum and C is minimum, and the value range of [ C, A ] is used as the prediction result.

The method can also comprise the following steps: and taking the average value of the single recovery factor value predicted by each target sub-model as the central value of the prediction range of the recovery factor of the oil reservoir, taking the mean square deviation of each single recovery factor value and the central value as the floating value of the prediction range of the recovery factor of the oil reservoir, and determining the numerical range obtained by increasing and decreasing the floating value of the central value as the prediction result of the recovery factor of the oil reservoir. For example, if the average value of the single values of the recovery factor predicted by the target sub-model is D and the mean square error corresponding to each single value of the recovery factor is f, the numerical range [ D-f, D + f ] is used as the prediction result.

Alternatively, the data format of the predicted recovery factor of each target sub-model may be a numerical range, and the data format of the reservoir recovery factor prediction result may be a numerical range. Determining the oil reservoir recovery rate prediction result according to the recovery rate predicted by each target sub-model, wherein the method specifically comprises the following steps:

and respectively determining a minimum value and a maximum value from the numerical range of the recovery factor predicted by each target sub-model, and determining the numerical range between the minimum value and the maximum value as the prediction result of the recovery factor of the oil reservoir. For example, the predicted recovery ratio ranges of the target sub-model are [ G, H ], [ I, J ] and [ K, L ], respectively, wherein the maximum value is J, and the minimum value is G, and the value range [ G, J ] is used as the prediction result.

FIG. 5 is a schematic flow diagram of a reservoir recovery prediction method, according to an embodiment. The embodiment relates to a specific process for automatically predicting the recovery ratio of a reservoir by a computer device. As shown in fig. 5, includes:

s51, obtaining characteristic parameters to be predicted; the characteristic parameters to be predicted comprise static parameters representing fixed characteristics of the oil reservoir and dynamic parameters representing variation characteristics of the oil reservoir.

Specifically, the computer equipment obtains characteristic parameters to be predicted, wherein the characteristic parameters to be predicted are related parameters of an oil reservoir needing recovery rate prediction. It should be noted that the predicted characteristic parameters include dynamic data and static data, and specific descriptions of the dynamic data and the static data may be referred to in the foregoing, and are not described herein again.

S52, processing the characteristic parameters to be predicted by adopting the oil reservoir recovery model according to any one of the embodiments, and outputting oil reservoir recovery prediction results; the oil reservoir recovery model comprises a plurality of target sub-models, and the oil reservoir recovery prediction result is determined according to the recovery predicted by each target sub-model.

Specifically, the computer device inputs the characteristic parameters to be predicted into a plurality of target sub-models in the oil reservoir recovery rate model, each target sub-model outputs a corresponding recovery rate, and the computer device determines an oil reservoir recovery rate prediction result according to the corresponding recovery rate output by each target sub-model.

And the computer equipment determines the oil reservoir recovery rate prediction result according to the distribution of the recovery rates output by the plurality of target sub-models. Optionally, the data format of the predicted recovery ratio of each target sub-model is a single value, the data format of the oil reservoir recovery ratio prediction result is a numerical range, the computer device respectively determines a minimum value and a maximum value from the single recovery ratio value predicted by each target sub-model, and the numerical range between the minimum value and the maximum value is determined as the oil reservoir recovery ratio prediction result; the computer device may also determine a minimum value and a maximum value from the predicted recovery factor value ranges for each of the target sub-models, respectively, and determine a value range between the minimum value and the maximum value as the predicted reservoir recovery factor. Optionally, the data format of the recovery factor predicted by each target sub-model is a numerical range, the data format of the reservoir recovery factor prediction result is a numerical range, the computer device may use an average value of the recovery factor single values predicted by each target sub-model as a central value of the reservoir recovery factor prediction range, use a mean square error of each predicted recovery factor single value and the central value as a floating value of the reservoir recovery factor prediction range, and determine the numerical range in which the central value increases and decreases the floating value as the reservoir recovery factor prediction result. The oil reservoir recovery rate prediction result obtained by the method can be calculated based on a plurality of accurate target sub-models, and compared with a single value of the traditional recovery rate, the oil reservoir recovery rate prediction result is in a numerical range, so that the oil reservoir recovery rate prediction result is more matched with the actual condition, and is more reasonable and accurate.

In this embodiment, the computer device uses a plurality of target sub-models in the oil reservoir recovery model in any of the above embodiments to process the characteristic parameters to be predicted respectively, so as to obtain an oil reservoir recovery prediction result. Because the plurality of target submodels included in the oil reservoir recovery ratio prediction model are the target submodels with the optimal test results, namely the target submodels with higher precision, which are obtained by screening, and the oil reservoir recovery ratio prediction results are determined according to the recovery ratio predicted by each target submodel, the technical problem of single application scene caused by adopting a single model to perform recovery ratio prediction in the traditional method can be avoided, so that the application scenes of the model are greatly enriched, and the robustness of the model is improved.

It should be understood that although the various steps in the flow charts of fig. 2-5 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-5 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternating with other steps or at least some of the sub-steps or stages of other steps.

In one embodiment, as shown in fig. 6, there is provided a reservoir recovery prediction model acquisition device, including:

an obtaining module 100, configured to obtain training feature parameters and testing feature parameters; the training characteristic parameters and the testing characteristic parameters comprise static parameters representing fixed characteristics of the oil deposit and dynamic parameters representing variation characteristics of the oil deposit;

the training module 200 is configured to respectively train a plurality of submodels to be trained by using the training characteristic parameters to obtain a plurality of submodels to be verified; the sub-models to be verified are models based on different algorithms, and each sub-model to be trained is a sub-model matched with the type of the training characteristic parameter;

the test module 300 is configured to input the test characteristic parameters into each to-be-verified sub-model respectively to obtain a test result of each to-be-verified sub-model;

the processing module 400 is used for screening a plurality of target sub-models from the plurality of sub-models to be verified according to the test result of each sub-model to be verified and preset screening requirements to obtain the plurality of target sub-models as the oil reservoir recovery rate prediction models;

In an embodiment, the obtaining module 100 is specifically configured to obtain an initial feature parameter; wherein the initial characteristic parameters comprise static parameters and dynamic parameters; carrying out data cleaning on the initial characteristic parameters to obtain intermediate characteristic parameters; splitting the intermediate characteristic parameters to obtain the training characteristic parameters and the testing characteristic parameters; the training characteristic parameters and the testing characteristic parameters comprise static parameters and dynamic parameters.

In an embodiment, the preset screening requirement is a preset number, and the processing module 400 is configured to sort each test result according to the error rate, and screen a preset number of submodels to be verified from the error rate sort as the target submodels.

In one embodiment, the training module 200 is specifically configured to obtain a plurality of initial predictor models; the initial prediction sub-model is a model for predicting the recovery ratio of the oil reservoir based on the characteristic parameters of different data types; and screening out a plurality of submodels to be trained from the plurality of initial prediction submodels according to the data types of the training characteristic parameters and the testing characteristic parameters.

In one embodiment, the data format of the recovery factor predicted by each target sub-model is a single value or a numerical range, and the data format of the reservoir recovery factor prediction result is a numerical range;

when the data format of the predicted recovery factor of each target sub-model is a single value, the processing module 400 is further configured to determine a minimum value and a maximum value from the single recovery factor predicted by each target sub-model, respectively, and determine a numerical range between the minimum value and the maximum value as the prediction result of the recovery factor of the oil reservoir;

the processing module 400 is further configured to use an average value of the single recovery factor values predicted by each target sub-model as a central value of the prediction range of the recovery factor of the oil reservoir, use a mean square error between each single recovery factor value predicted by each target sub-model and the central value as a floating value of the prediction range of the recovery factor of the oil reservoir, and determine a numerical range obtained by increasing and decreasing the floating value of the central value as a prediction result of the recovery factor of the oil reservoir;

in one embodiment, when the data format of the predicted recovery factor of each of the target sub-models is a numerical range, the processing module 400 is further configured to determine a minimum value and a maximum value from the predicted recovery factor numerical range of each of the target sub-models, respectively, and determine the numerical range between the minimum value and the maximum value as the prediction result of the recovery factor of the reservoir.

In one embodiment, as shown in fig. 7, there is provided a reservoir recovery prediction device comprising:

an obtaining module 500, configured to obtain a feature parameter to be predicted; the characteristic parameters to be predicted comprise static parameters representing fixed characteristics of the oil reservoir and dynamic parameters representing variation characteristics of the oil reservoir;

the prediction module 600 is configured to process the characteristic parameter to be predicted in the oil reservoir recovery model according to any of the embodiments, and output an oil reservoir recovery prediction result; the oil reservoir recovery model comprises a plurality of target sub-models, and the oil reservoir recovery prediction result is determined according to the recovery predicted by each target sub-model.

The oil reservoir recovery ratio prediction model acquisition device comprises a processor and a memory, wherein the acquisition module 100, the training module 200, the testing module 300, the processing module 400 and the like are stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions. The oil reservoir recovery prediction device comprises a processor and a memory, wherein the acquisition module 500, the prediction module 600 and the like are stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.

The processor comprises a kernel, and the kernel calls the corresponding program unit from the memory. One or more than one kernel can be set, and the accuracy and the rationality of the oil reservoir recovery ratio prediction are improved by adjusting the kernel parameters.

An embodiment of the present invention provides a storage medium having a program stored thereon, which when executed by a processor, implements the method for obtaining a reservoir recovery prediction model.

An embodiment of the present invention provides a storage medium having a program stored thereon, which when executed by a processor implements the reservoir recovery prediction method.

Embodiments of the present invention provide a processor for executing a program, wherein the program executes the method for obtaining a reservoir recovery prediction model during execution.

Embodiments of the invention provide a processor for running a program, wherein the program is run to perform the reservoir recovery prediction method.

The embodiment of the invention provides equipment, which comprises at least one processor, at least one memory and a bus, wherein the memory and the bus are connected with the processor; the processor and the memory complete mutual communication through a bus; the processor is configured to invoke program instructions in the memory to perform the reservoir recovery prediction model acquisition method described above. The device herein may be a server, a PC, a PAD, a mobile phone, etc.

The embodiment of the invention provides equipment, which comprises at least one processor, at least one memory and a bus, wherein the memory and the bus are connected with the processor; the processor and the memory complete mutual communication through a bus; the processor is configured to invoke program instructions in the memory to perform the reservoir recovery prediction method described above. The device herein may be a server, a PC, a PAD, a mobile phone, etc.

The present application further provides a computer program product adapted to perform a program for initializing the following method steps when executed on a data processing device:

In an embodiment, the program, when being executed on a data processing device, is further adapted to perform a procedure for initializing the following method steps:

acquiring initial characteristic parameters; wherein the initial characteristic parameters comprise static parameters and dynamic parameters;

carrying out data cleaning on the initial characteristic parameters to obtain intermediate characteristic parameters;

splitting the intermediate characteristic parameters to obtain the training characteristic parameters and the testing characteristic parameters; the training characteristic parameters and the testing characteristic parameters comprise static parameters and dynamic parameters.

In an embodiment, the predetermined filtering requirement is a predetermined number, which when executed on the data processing device is further adapted to execute a program initializing the following method steps:

sequencing each test result according to the error rate;

screening a preset number of submodels to be verified from the error rate sequence to be used as the target submodels;

and/or the presence of a gas in the gas,

the training of the plurality of submodels to be trained by adopting the training characteristic parameters comprises the following steps of:

acquiring a plurality of initial predictor models; the initial prediction sub-model is a model for predicting the recovery ratio of the oil reservoir based on the characteristic parameters of different data types;

and screening out a plurality of submodels to be trained from the plurality of initial prediction submodels according to the data types of the training characteristic parameters and the testing characteristic parameters.

In one embodiment, the data format of the recovery factor predicted by each target sub-model is a single value or a numerical range, and the data format of the reservoir recovery factor prediction result is a numerical range; when executed on a data processing device, is further adapted to perform a procedure for initializing the following method steps:

and/or the presence of a gas in the gas,

taking the average value of the single recovery factor values predicted by each target sub-model as the central value of the prediction range of the recovery factor of the oil reservoir, taking the mean square deviation of each single recovery factor value and the central value as the floating value of the prediction range of the recovery factor of the oil reservoir, and determining the numerical range obtained by increasing and decreasing the floating value of the central value as the prediction result of the recovery factor of the oil reservoir;

and/or the presence of a gas in the gas,

and respectively determining a minimum value and a maximum value from the numerical range of the recovery factor predicted by each target sub-model, and determining the numerical range between the minimum value and the maximum value as the prediction result of the recovery factor of the oil reservoir.

In one embodiment, the data format of the recovery factor predicted by each of the target sub-models is a single value or a range of values, and the data format of the reservoir recovery factor prediction is a range of values, and when executed on the data processing device is further adapted to perform a procedure which initializes a procedure having the following method steps:

when the data format of the predicted recovery ratio of each target sub-model is a single value, the oil reservoir recovery ratio prediction result is determined according to the predicted recovery ratio of each target sub-model, and the method comprises the following steps:

and/or the presence of a gas in the gas,

when the data format of the predicted recovery ratio of each target sub-model is a numerical range, the oil reservoir recovery ratio prediction result is determined according to the predicted recovery ratio of each target sub-model, and the method comprises the following steps:

and/or the presence of a gas in the gas,

In one embodiment, there is provided a computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the method for obtaining a reservoir recovery prediction model in any of the above embodiments when executing the computer program, specifically comprising the steps of:

It should be clear that, in the embodiments of the present application, the process of executing the computer program by the processor is consistent with the process of executing the steps in the above method, and specific reference may be made to the description above.

In one embodiment, there is provided a computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the reservoir recovery prediction method in any of the above embodiments when executing the computer program, specifically comprising:

the oil reservoir recovery rate model comprises a plurality of target sub-models, the oil reservoir recovery rate prediction result is determined according to the recovery rate predicted by each target sub-model, and the oil reservoir recovery rate prediction range is determined according to a plurality of single predicted recovery rate values.

In one embodiment, there is provided a computer readable storage medium having a computer program stored thereon, the computer program when executed by a processor implementing the method for reservoir recovery predictive model acquisition in any of the above embodiments, comprising:

In one embodiment, there is provided a computer readable storage medium having a computer program stored thereon, the computer program when executed by a processor implementing the method for reservoir recovery prediction in any of the above embodiments, comprising:

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a device includes one or more processors (CPUs), memory, and a bus. The device may also include input/output interfaces, network interfaces, and the like.

The memory may include volatile memory in a computer readable medium, Random Access Memory (RAM) and/or nonvolatile memory such as Read Only Memory (ROM) or flash memory (flash RAM), and the memory includes at least one memory chip. The memory is an example of a computer-readable medium.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, 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 disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in the process, method, article, or apparatus that comprises the element.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The above are merely examples of the present application and are not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for obtaining a reservoir recovery prediction model, the method comprising:

2. The method of claim 1, wherein the obtaining training feature parameters and testing feature parameters comprises:

3. The method according to claim 1 or 2, wherein the preset screening requirements are a preset number, and the screening of the plurality of target submodels from the plurality of submodels to be verified according to the screening requirements and the test result of each submodel to be verified comprises:

sequencing each test result according to the error rate;

and/or the presence of a gas in the gas,

4. The method of claim 1, wherein:

the data format of the recovery factor predicted by each target sub-model is a single value or a numerical range, and the data format of the oil reservoir recovery factor prediction result is a numerical range;

and/or the presence of a gas in the gas,

5. A method for reservoir recovery prediction, the method comprising:

processing the characteristic parameter to be predicted by using the reservoir recovery model according to any one of claims 1 to 4 to output a reservoir recovery prediction result;

6. The method of claim 5, wherein the data format for the recovery factor predicted by each of the target sub-models is a single value or a numerical range, and the data format for the reservoir recovery factor predicted result is a numerical range;

and/or the presence of a gas in the gas,

7. An apparatus for obtaining a reservoir recovery prediction model, the apparatus comprising:

8. A reservoir recovery prediction device, the device comprising:

a prediction module, for processing the characteristic parameter to be predicted in the reservoir recovery model according to any one of claims 1 to 4, and outputting a reservoir recovery prediction result; the oil reservoir recovery model comprises a plurality of target sub-models, and the oil reservoir recovery prediction result is determined according to the recovery predicted by each target sub-model.

9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 6 when executing the computer program.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.