Method and device for determining oil-water two-phase relative permeability curve based on machine learning

Abstract

The invention provides a method and device for determining an oil-water two-phase relative permeability curve based on machine learning, and the method comprises the steps: obtaining a plurality of groups of target parameters of a developed reservoir, and determining a first reference undetermined coefficient in an oil-phase relative permeability curve model and a second reference undetermined coefficient in a water-phase relative permeability curve model of the developed reservoir; training a machine learning model based on a sample data set constructed by the target parameter, the first reference to-be-determined coefficient and the second reference to-be-determined coefficient; predicting an oil-water two-phase relative permeability curve of the target oil reservoir by using the qualified trained machine learning model to obtain the oil-water two-phase relative permeability curve; therefore, when the oil-water two-phase relative permeability curve of the target oil reservoir needs to be determined, only the rock parameters and the fluid parameters of the target oil reservoir need to be obtained, the undetermined coefficients corresponding to the oil-water two-phase relative permeability curve model are output by utilizing the qualified trained machine learning model based on the rock parameters and the fluid parameters, and an oil-water relative permeability experiment does not need to be newly added; and the oil-water two-phase permeability curve can be rapidly and accurately output.

Description

Technical Field

[0001] The present application relates to the technical field of oil and gas development, and in particular to a method, device, medium and equipment for determining an oil-water two-phase permeability curve. Background Art

[0002] The relative permeability curve reflects the relationship between the relative permeability of a phase (oil phase or water phase) and its saturation. Relative permeability is the ratio of the effective permeability of a phase of fluid to the absolute permeability of the reservoir when multiphase fluids coexist and flow in the reservoir, reflecting the relative flow capacity of each phase of fluid when it seeps in the reservoir.

[0003] The oil-water two-phase relative permeability curve is the basis for studying oil-water two-phase seepage. It can comprehensively reflect the oil-water two-phase seepage characteristics and reservoir properties. It is an indispensable and important data for oilfield development parameter calculation, dynamic analysis and reservoir numerical simulation research. It can be used in historical matching and subsequent development plan prediction process, and guide on-site construction operations in actual production.

[0004] In practical applications, there are many factors that affect the oil-water two-phase relative permeability curve. For new research areas, additional phase permeability experiments are needed to determine the oil-water two-phase relative permeability curve. The determination process is rather cumbersome and the workload is large.

[0005] Therefore, how to achieve rapid and accurate prediction of the oil-water two-phase permeability curve in the new research area is a technical problem that needs to be solved at present. Summary of the invention

[0006] In response to the problems existing in the prior art, the embodiments of the present invention provide a method, device, medium and equipment based on determining the oil-water two-phase relative permeability curve, so as to solve or partially solve the technical problem that when determining the oil-water two-phase relative permeability curve of a new research area in the prior art, it is necessary to add an oil-water relative permeability experiment, the process is cumbersome and the workload is large, resulting in the inability to accurately and quickly determine the oil-water two-phase relative permeability curve of the new research area.

[0007] In a first aspect of the present invention, a method for determining an oil-water two-phase permeability curve based on machine learning is provided, the method comprising:

[0008] Acquire multiple sets of target parameters of developed reservoirs, and determine the first reference coefficients to be determined in the oil phase permeability curve model and the second reference coefficients to be determined in the water phase permeability curve model of the developed reservoirs; the target parameters include: rock parameters and fluid parameters;

[0009] Constructing a sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined;

[0010] Constructing a machine learning model, and training the machine learning model according to the sample data set to obtain a trained machine learning model;

[0011] The trained and qualified machine learning model is used to predict the oil-water two-phase permeability curve of the target reservoir to obtain the oil-water two-phase permeability curve.

[0012] In the above scheme, the step of determining the first reference coefficient to be determined in the oil phase relative permeability curve model and the second reference coefficient to be determined in the water phase relative permeability curve model of the developed reservoir includes:

[0013] For any set of target parameters, at different water saturations, a phase permeability curve simulation is performed based on the target parameters to obtain a corresponding oil phase phase permeability curve and a water phase phase permeability curve;

[0014] Fitting a first function corresponding to the oil phase permeability curve and a second function corresponding to the water phase permeability curve;

[0015] The first reference coefficient to be determined is extracted from the first function, and the second reference coefficient to be determined is extracted from the second function.

[0016] In the above scheme, the first function is

[0017] The second function is:

[0018] The first reference coefficient to be determined includes a and b; the second reference coefficient to be determined includes c and d; the K ro is the oil phase relative permeability, the K o (S wi ) is the oil phase relative permeability corresponding to the irreducible water saturation, and the S wd is the dimensionless water saturation; rw is the relative permeability of water phase.

[0019] In the above solution, before constructing the sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined, the method further includes:

[0020] According to the formula Normalize the target parameters respectively to obtain the normalized target parameters

[0021] If it is determined that the first reference coefficient to be determined and / or the second reference coefficient to be determined need to be processed with the same order of magnitude, the first reference coefficient to be determined and / or the second reference coefficient to be determined are reduced by a target multiple so that the values ​​of the first reference coefficient to be determined and the second reference coefficient to be determined are between [0, 1]; wherein,

[0022] The z max is the maximum value of the target parameter before normalization, z min is the minimum value of the target parameter before normalization, and z is any target parameter before normalization.

[0023] In the above solution, the step of constructing a sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined includes:

[0024] Taking each set of the target parameters as a reference data input set, and taking the corresponding first reference coefficient to be determined and the second reference coefficient to be determined as a reference data output set;

[0025] The sample data set is constructed according to the reference data input set and the reference data output set.

[0026] In the above solution, the machine learning model is trained according to the sample data set to obtain a qualified machine learning model, including:

[0027] Dividing the sample data set into a training set and a test set;

[0028] The machine learning model is trained using the training set, and when the trained machine learning model is tested using the test set, if it is determined that the test accuracy meets the requirements, the corresponding machine learning model is output.

[0029] In the above scheme, the machine learning model is a deep neural network model, which includes an input layer, at least 2 hidden layers and an output layer; the deep neural network model is:

[0030]

[0031] The k is the output value of the kth layer of the deep neural network, and the output value of the output layer is the first undetermined coefficient and the second undetermined coefficient; x i is the input value of the i-th layer. When the i-th layer is the input layer, x i is the target parameter of the target reservoir; f o is the activation function of the output layer, f h is the activation function of the hidden layer, i is the number of the input layer neuron, j is the number of the hidden layer neuron, k is the number of the output layer neuron, N i is the number of neurons in the input layer, N h is the number of neurons in the output layer, W ij represents the connection weight between the i-th input layer neuron and the j-th hidden layer neuron, W jkrepresents the connection weight between the jth hidden layer neuron and the kth output layer neuron, (b h ) j represents the bias of the jth hidden layer neuron, (b o ) k Represents the bias of the kth output layer neuron.

[0032] In the above scheme, the method of using the trained machine learning model to predict the oil-water two-phase permeability curve of the target reservoir includes:

[0033] Using the target parameter of the target reservoir as an input value of the trained machine learning model, and using the trained machine learning model to output a first current undetermined coefficient and a second current undetermined coefficient;

[0034] Determining an oil phase relative permeability curve model of the target oil reservoir according to the first currently undetermined coefficient, and determining a water phase relative permeability curve model of the target oil reservoir according to the second currently undetermined coefficient;

[0035] The oil-water two-phase permeability curve of the target oil reservoir is determined according to the oil phase permeability curve model of the target oil reservoir and the water phase permeability curve model of the target oil reservoir.

[0036] A second aspect of the present invention provides a device for determining an oil-water two-phase permeability curve based on machine learning, the device comprising:

[0037] A determination unit is used to obtain multiple groups of target parameters of developed reservoirs, and determine the first reference coefficient to be determined in the oil phase permeability curve model and the second reference coefficient to be determined in the water phase permeability curve model of the developed reservoirs; the target parameters include: rock parameters and fluid parameters;

[0038] A construction unit, configured to construct a sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined;

[0039] A training unit, used to construct a machine learning model, train the machine learning model according to the sample data set, and obtain a qualified machine learning model;

[0040] The prediction unit is used to predict the oil-water two-phase permeability curve of the target reservoir by using the trained machine learning model to obtain the oil-water two-phase permeability curve.

[0041] According to a third aspect of the present invention, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any one of the methods described in the first aspect are implemented.

[0042] The present invention provides a method and device for determining an oil-water two-phase permeability curve based on machine learning, the method comprising: obtaining multiple groups of target parameters of developed reservoirs, and determining a first reference to-be-determined coefficient in an oil phase permeability curve model and a second reference to-be-determined coefficient in a water phase permeability curve model of the developed reservoirs; the target parameters comprising: rock parameters and fluid parameters; constructing a sample data set according to the target parameters and the first reference to-be-determined coefficient and the second reference to-be-determined coefficient; constructing a machine learning model, training the machine learning model according to the sample data set, and obtaining a machine learning model that has passed the training; using the machine learning model that has passed the training to predict the oil-water two-phase permeability curve of a target reservoir, and obtaining the oil-water two-phase permeability curve; in this way, when it is necessary to determine the oil-water two-phase permeability curve of a target reservoir, it is only necessary to obtain the rock parameters and fluid parameters of the target reservoir, and based on the rock parameters and the fluid parameters, the to-be-determined coefficients corresponding to the oil-water two-phase permeability curve model are output using the trained machine learning model, without the need for adding an oil-water permeability experiment, and the oil-water two-phase permeability curve can be output quickly and accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] By reading the detailed description of the preferred embodiment below, various other advantages and benefits will become clear to those of ordinary skill in the art. The accompanying drawings are only used for the purpose of illustrating the preferred embodiment and are not considered to be limitations of the present invention. In addition, the same reference symbols are used to represent the same components throughout the accompanying drawings.

[0044] In the attached picture:

[0045] Figure 1 A schematic flow chart of a method for determining an oil-water two-phase permeability curve based on machine learning according to an embodiment of the present invention is shown;

[0046] Figure 2 A logical schematic diagram of constructing a machine learning model according to an embodiment of the present invention is shown;

[0047] Figure 3 A schematic diagram of the structure of a device for determining an oil-water two-phase permeability curve based on machine learning according to an embodiment of the present invention is shown;

[0048] Figure 4 A schematic diagram of a reference oil phase permeability curve and a reference water phase permeability curve obtained by performing a phase permeability curve simulation experiment on the rock parameters and fluid parameters in Table 1 according to an embodiment of the present invention is shown;

[0049] Figure 5 A schematic diagram of a network topology structure of a machine learning model according to an embodiment of the present invention is shown;

[0050] Figure 6A schematic diagram of a curve showing a change in error value during an iteration of a machine learning model according to an embodiment of the present invention is shown;

[0051] Figure 7 A schematic diagram showing a comparison between the predicted value of oil phase relative permeability predicted based on a training set and a test set respectively using a machine learning model according to an embodiment of the present invention and the actual value;

[0052] Figure 8 A schematic diagram showing a comparison between the predicted value and the actual value of the relative permeability of the water phase predicted based on the training set and the test set respectively by using a machine learning model according to an embodiment of the present invention;

[0053] Fig. 9 A schematic diagram of an oil-water two-phase relative permeability curve of a target oil reservoir predicted by a machine learning model according to an embodiment of the present invention is shown;

[0054] Fig.10 A schematic diagram of an oil-water two-phase relative permeability curve of another target oil reservoir predicted by a machine learning model according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0055] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0056] The present invention provides a method for determining the oil-water two-phase permeability curve based on machine learning, such as Figure 1 As shown, the method comprises the following steps:

[0057] S110, obtaining multiple groups of target parameters of developed reservoirs, and determining a first reference coefficient to be determined in an oil phase permeability curve model and a second reference coefficient to be determined in a water phase permeability curve model of the developed reservoirs; the target parameters include: rock parameters and fluid parameters;

[0058] In the embodiments of this specification, the target parameters include rock parameters and fluid parameters. The rock parameters include rock properties, porosity, absolute permeability, wettability, initial water saturation and residual oil saturation; the fluid parameters include crude oil viscosity and water viscosity. In this embodiment, multiple sets of target parameters can be collected from the field data of the developed oil reservoir.

[0059] It can be seen that each group of target parameters includes 8. When conducting phase permeability simulation experiments based on target parameters, multiple groups of target parameters can be obtained, and the values ​​of each group of target parameters can be different. Then, a phase permeability curve simulation experiment is carried out based on each group of target parameters to obtain the corresponding oil phase phase permeability curve model (function) of the developed reservoir, and then the first reference unknown coefficient in the oil phase phase permeability curve model and the second reference unknown coefficient in the water phase phase permeability curve model are extracted.

[0060] It can be understood that, for each set of target parameters, there exists an oil phase permeability curve model and a water phase permeability curve model corresponding to each set of target parameters.

[0061] Then, in one embodiment, determining the first reference coefficient to be determined in the oil phase relative permeability curve model of the developed reservoir and the second reference coefficient to be determined in the water phase relative permeability curve model includes:

[0062] For any set of target parameters, at different water saturations, a phase permeability curve simulation is performed based on the target parameters to obtain the corresponding reference oil phase phase permeability curve and reference water phase phase permeability curve;

[0063] Fitting a first function corresponding to a reference oil phase permeability curve and a second function corresponding to a reference water phase permeability curve;

[0064] A first reference coefficient to be determined is extracted from the first function, and a second reference coefficient to be determined is extracted from the second function.

[0065] The process of fitting the first function according to the reference oil phase permeability curve and the process of fitting the second function according to the reference water phase permeability curve are equivalent to the process of solving the de-curve.

[0066] In one embodiment, the first function is The second function is:

[0067] The first reference coefficients to be determined include a and b; the second reference coefficients to be determined include c and d; K ro is the relative permeability of the oil phase, K o (S wi ) is the oil phase relative permeability corresponding to irreducible water saturation, S wd is the dimensionless water saturation; K rw is the relative permeability of water phase.

[0068] That is to say, after conducting a permeability curve simulation experiment on any set of target parameters, the corresponding reference oil phase permeability curve and reference water phase permeability curve will be output. From a mathematical point of view, each curve can be characterized by a function. Therefore, after obtaining the reference oil phase permeability curve and the reference water phase permeability curve, the oil phase permeability curve can be characterized by the first function, and the water phase permeability curve can be characterized by the second function.

[0069] In the process of performing phase permeability curve simulation experiment for any set of target parameters, the simulation tool can be directly used to output the first function corresponding to the oil phase phase permeability curve, and the simulation tool can be used to output the second function corresponding to the water phase phase permeability curve. Then, the corresponding first reference coefficients a and b are extracted from the first function, and the corresponding second reference coefficients c and d are extracted from the second function. In this way, each set of target parameters corresponds to a set of a, b, c, and d values.

[0070] S111, constructing a sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined;

[0071] After the first reference coefficients and second reference coefficients corresponding to each set of target parameters are determined, a sample data set of the machine learning model is constructed based on the target parameters and the first reference coefficients and second reference coefficients.

[0072] In practical applications, there may be some bad pixel data. Therefore, in order to improve the accuracy of subsequent machine learning models, it is necessary to first process the target parameters, the first reference coefficients to be determined, and the second reference coefficients to be determined. Then, use the processed target parameters, the first reference coefficients to be determined, and the second reference coefficients to construct a sample data set.

[0073] In one implementation, constructing a sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined includes:

[0074] According to the formula Normalize the target parameters respectively to obtain the normalized target parameters

[0075] If it is determined that the first reference coefficient to be determined and / or the second reference coefficient to be determined need to be processed with the same order of magnitude, the first reference coefficient to be determined and / or the second reference coefficient to be determined are reduced by a target multiple so that the values ​​of the first reference coefficient to be determined and the second reference coefficient to be determined are between [0,1]; wherein,

[0076] z max is the maximum value of the target parameter before normalization, z min is the minimum value of the target parameter before normalization, and z is any target parameter before normalization.

[0077] Since the data properties of the target parameters and the coefficients to be determined are different, the processing methods for the target parameters and the coefficients to be determined are also different. For the target parameters, the above normalization function can be used to normalize each target parameter.

[0078] For example, if the porosity in the target parameters is normalized, the maximum porosity and the minimum porosity need to be determined from all groups of target parameters first; and then the porosity in each group of target parameters is normalized using the above normalization formula.

[0079] Assuming that the a value of the first undetermined coefficient is processed, then it is necessary to first determine whether the a value needs to be processed. If the a value is much larger than the values ​​of the remaining undetermined coefficients (for example, the a value is about 10 times the value of the remaining undetermined coefficients), then the a value needs to be reduced by a target multiple (10 times), where the target multiple is determined based on the order of magnitude of the a value and the order of magnitude of the remaining undetermined coefficients.

[0080] Then, a sample data set is constructed using the processed target parameters, the first reference coefficients to be determined, and the second reference coefficients to be determined, including:

[0081] Taking each set of target parameters as a reference data input set, and taking the corresponding first reference coefficient to be determined and the second reference coefficient to be determined as a reference data output set;

[0082] Construct a sample data set based on a reference data input set and a reference data output set.

[0083] That is, a set of sample data includes target parameters and corresponding first and second undetermined coefficients.

[0084] S112, constructing a machine learning model, and training the machine learning model according to the sample data set to obtain a trained machine learning model;

[0085] In the embodiments of this specification, the machine learning model is a deep neural network model, which includes an input layer, at least two hidden layers and an output layer; the deep neural network model is:

[0086]

[0087] y k is the output value of the kth layer of the deep neural network, and the output value of the output layer is the first undetermined coefficient and the second undetermined coefficient; i is the input value of the i-th layer. When the i-th layer is the input layer, x i is the target parameter of the target reservoir; f o is the activation function of the output layer, f h is the activation function of the hidden layer, i is the number of the input layer neuron, j is the number of the hidden layer neuron, k is the number of the output layer neuron, N i is the number of neurons in the input layer, N h is the number of neurons in the output layer, W ijrepresents the connection weight between the i-th input layer neuron and the j-th hidden layer neuron, W jk represents the connection weight between the jth hidden layer neuron and the kth output layer neuron, (b h ) j represents the bias of the jth hidden layer neuron, (b o ) k Represents the bias of the kth output layer neuron.

[0088] refer to Figure 2 , the data set of the input layer of the neural network in this specification is rock parameters and fluid parameters, and the data set of the output layer is the first unknown coefficient and the second unknown coefficient.

[0089] After the machine learning model is constructed, in one embodiment, the machine learning model is trained according to the sample data set to obtain a trained machine learning model, including:

[0090] Divide the sample data set into training set and test set;

[0091] The machine learning model is trained using the training set, and when the trained machine learning model is tested using the test set, if it is determined that the test accuracy meets the requirements, the corresponding machine learning model is output.

[0092] If it is determined that the test accuracy does not meet the requirements, it is necessary to adjust the learning parameters of the machine learning model, such as learning rate, number of iterations, etc., and then train and test the adjusted machine learning model again until the test accuracy meets the requirements and output the corresponding machine learning model.

[0093] S113, using the trained machine learning model to predict the oil-water two-phase permeability curve of the target reservoir to obtain the oil-water two-phase permeability curve.

[0094] In one embodiment, the oil-water two-phase permeability curve of the target reservoir is predicted using a trained machine learning model, including:

[0095] Using the target parameters of the target reservoir as input values ​​of the trained machine learning model, and using the trained machine learning model to output a first current undetermined coefficient and a second current undetermined coefficient;

[0096] Determining an oil phase relative permeability curve model of a target oil reservoir according to a first currently undetermined coefficient, and determining a water phase relative permeability curve model of the target oil reservoir according to a second currently undetermined coefficient;

[0097] The oil-water two-phase permeability curve of the target reservoir is determined according to the oil phase permeability curve model of the target reservoir and the water phase permeability curve model of the target reservoir.

[0098] Specifically, if the oil-water two-phase permeability curve of the target reservoir needs to be predicted, the rock parameters and fluid parameters of the target reservoir need to be obtained, and the rock parameters and fluid parameters are used as the input values ​​of the machine learning model. Then the output of the machine learning model is the first undetermined coefficients a and b and the second undetermined coefficients c and d;

[0099] Then, the first unknown coefficient is substituted into the first function corresponding to the oil phase relative permeability curve, and the oil phase relative permeability curve of the target oil reservoir is drawn according to the first function; the second unknown coefficient is substituted into the second function corresponding to the water phase relative permeability curve, and the water phase relative permeability curve of the target oil reservoir is drawn according to the second function.

[0100] It can be seen that when it is necessary to determine the oil-water two-phase permeability curve of the target oil reservoir, it is only necessary to obtain the rock parameters and fluid parameters of the target oil reservoir, and use the trained qualified machine learning model based on the rock parameters and fluid parameters to output the undetermined coefficients corresponding to the oil-water two-phase permeability curve model. There is no need to add new oil-water permeability experiments, and the oil-water two-phase permeability curve can be output quickly and accurately.

[0101] Based on the same inventive concept as in the above-mentioned embodiment, the embodiment of this specification also provides a device for determining the oil-water two-phase permeability curve based on machine learning, such as Figure 3 The device comprises:

[0102] The determination unit 31 is used to obtain multiple groups of target parameters of developed reservoirs, and determine the first reference coefficient to be determined in the oil phase permeability curve model and the second reference coefficient to be determined in the water phase permeability curve model of the developed reservoir; the target parameters include: rock parameters and fluid parameters;

[0103] A construction unit 32, configured to construct a sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined;

[0104] A training unit 33 is used to construct a machine learning model, and train the machine learning model according to the sample data set to obtain a trained machine learning model;

[0105] The prediction unit 34 is used to predict the oil-water two-phase permeability curve of the target reservoir by using the trained machine learning model to obtain the oil-water two-phase permeability curve.

[0106] Since the device introduced in the embodiment of the present invention is a device used to implement the method for determining the oil-water two-phase permeability curve based on machine learning in the embodiment of the present invention, based on the method introduced in the embodiment of the present invention, the person skilled in the art can understand the specific structure and deformation of the device, so it is not repeated here. All devices used in the method of the embodiment of the present invention belong to the scope of protection of the present invention.

[0107] Based on the same inventive concept, this embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any step of the method described above when executing the computer program.

[0108] Based on the same inventive concept, this embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the methods described above are implemented.

[0109] In practical applications, when the method and device for determining the oil-water two-phase permeability curve based on machine learning provided in the above embodiment are used to predict the oil-water two-phase permeability curve of the target reservoir, the following is achieved:

[0110] First, the rock parameters and fluid parameters of the developed reservoir are collected, and the parameter collection table is shown in Table 1. Table 1 shows 8 types of input parameters collected from 50 groups of samples, including the unit, upper limit and lower limit of the parameters. Among them, the reservoir rock parameters include the lithology, porosity, absolute permeability, wettability, initial water saturation and residual oil saturation of the reservoir rock; the fluid parameters include crude oil viscosity and water viscosity.

[0111] Table 1

[0112] Input Parameters unit Lower limit Upper limit Lithology / Carbonate rock sandstone Porosity fraction 0.041 0.356 Absolute permeability mD 0.01 6500 Initial water saturation fraction 0.069 0.694 Residual oil saturation fraction 0.121 0.6 Wettability / Water Oil wet Crude oil viscosity mPa·s 1.2 1538 Water viscosity mPa·s 0.14 1.4

[0113] Then, the rock parameters and fluid parameters in Table 1 are subjected to phase permeability curve simulation experiments to obtain the corresponding reference oil phase permeability curve and reference water phase permeability curve; wherein the reference oil phase permeability curve and the reference water phase permeability curve are as follows: Figure 4 shown. Figure 4 The “·” in the figure refers to the reference oil phase permeability curve formed by fitting using simulation tools. It refers to the original oil phase permeability curve; It refers to the reference water phase relative permeability curve formed by fitting using simulation tools. It refers to the original water phase permeability curve. Figure 4 It can be seen that the reference oil-water two-phase permeability curve fitted by the simulation software in the embodiment of this specification has a high degree of conformity with the original oil-water two-phase permeability curve, and the correlation coefficient is above 0.98, indicating that the accuracy of the first and second undetermined coefficients determined based on the reference oil phase permeability curve and the reference water phase permeability curve in the embodiment of this specification can also be ensured.

[0114] The first undetermined coefficient a ranges from 3.482 to 10.93, the value range of b ranges from 0.5877 to 2.47, the value range of the second undetermined coefficient c ranges from 0.2717 to 3.737, and the value range of d ranges from -2.895 to 0.8744.

[0115] The rock parameters, fluid parameters, first reference coefficients to be determined and second reference coefficients to be determined are processed, and then a sample data set is constructed using the processed target parameters, first reference coefficients to be determined and second reference coefficients to be determined.

[0116] Specifically, the normalization function is used to normalize the rock parameters and fluid parameters. Since the a value is much larger than the value of the remaining coefficients to be determined (about 10 times the value of the remaining coefficients to be determined), the a value needs to be reduced by the target multiple (10 times).

[0117] A sample data set is constructed using the processed target parameters, the first reference coefficients to be determined, and the second reference coefficients to be determined.

[0118] Build a machine learning model. The machine learning model is a deep neural network. The network topology is as follows: Figure 5 As shown, the input layer has 8 neurons, including 8 input parameters; there are 3 hidden layers, and the number of neurons in the hidden layers are 21, 23, and 15 respectively; the output layer has 4 neurons, and outputs 2 first coefficients to be determined and 2 second coefficients to be determined.

[0119] Divide the sample data set into a training set and a test set; use the training set to train the machine learning model. If it is determined that the test accuracy does not meet the requirements, it is necessary to adjust the learning parameters of the machine learning model, such as adjusting the grid structure of the deep neural network, and at the same time improve the model accuracy through the Adam optimization algorithm, such as Figure 6 As shown in the figure, after a certain number of iterations, the error value of the entire model has been reduced to a lower level.

[0120] Furthermore, if Figure 7 As shown in the figure, when the training set and the test set are used as the test data respectively, the trained machine learning model is used to predict the training set and the test set to obtain the corresponding predicted oil phase relative permeability values. Compared with the actual values ​​of the oil phase relative permeability (known), the predicted oil phase relative permeability values ​​and the actual values ​​of the oil phase relative permeability are more evenly distributed near the 45° line.

[0121] Similarly, if Figure 8 As shown in the figure, when the training set and the test set are used as the test data, the trained machine learning model is used to predict the training set and the test set to obtain the corresponding water phase relative permeability prediction value. Compared with the actual value of the oil phase relative permeability (known), the water phase relative permeability prediction value and the actual value of the water phase relative permeability are more evenly distributed near the 45° line. This shows that the prediction accuracy of the machine learning model is high.

[0122] Then, the trained machine learning model can be used to predict the oil-water two-phase permeability curve of the target reservoir.

[0123] Specifically, it is necessary to obtain the rock parameters and fluid parameters of the target oil reservoir, and use the rock parameters and fluid parameters as input values ​​of the machine learning model. Then the output of the machine learning model is the first unknown coefficients a and b and the second unknown coefficients c and d.

[0124] It should be noted that, since the a value has been reduced in multiples, after obtaining the first undetermined coefficient a value, the a value needs to be restored, that is, the a value is multiplied by the target multiple to obtain the true value.

[0125] Then, the first unknown coefficient is substituted into the first function corresponding to the oil phase relative permeability curve, and the oil phase relative permeability curve of the target reservoir is plotted according to the first function; the second unknown coefficient is substituted into the second function corresponding to the water phase relative permeability curve, and the water phase relative permeability curve of the target reservoir is plotted according to the second function. Fig. 9 shown.

[0126] Similarly, when the machine learning model is used to predict the phase permeability curve of another target reservoir, the predicted oil-water two-phase phase permeability curve is as follows: Fig.10 shown.

[0127] Through one or more embodiments of the present invention, the present invention has the following beneficial effects or advantages:

[0128] The present invention provides a method and device for determining an oil-water two-phase permeability curve based on machine learning, the method comprising: obtaining multiple groups of target parameters of developed reservoirs, and determining a first reference to-be-determined coefficient in an oil phase permeability curve model and a second reference to-be-determined coefficient in a water phase permeability curve model of the developed reservoirs; the target parameters comprising: rock parameters and fluid parameters; constructing a sample data set according to the target parameters and the first reference to-be-determined coefficient and the second reference to-be-determined coefficient; constructing a machine learning model, training the machine learning model according to the sample data set, and obtaining a machine learning model that has passed the training; using the machine learning model that has passed the training to predict the oil-water two-phase permeability curve of a target reservoir, and obtaining the oil-water two-phase permeability curve; in this way, when it is necessary to determine the oil-water two-phase permeability curve of a target reservoir, it is only necessary to obtain the rock parameters and fluid parameters of the target reservoir, and based on the rock parameters and the fluid parameters, the to-be-determined coefficients corresponding to the oil-water two-phase permeability curve model are output using the trained machine learning model, without the need for adding an oil-water permeability experiment, and the oil-water two-phase permeability curve can be output quickly and accurately.

[0129] The algorithm and display provided herein are not inherently related to any particular computer, virtual system or other device. Various general purpose systems can also be used together with the teachings based on this. According to the above description, it is obvious that the structure required for constructing such systems. In addition, the present invention is not directed to any specific programming language either. It should be understood that various programming languages ​​can be utilized to realize the content of the present invention described herein, and the description of the above specific languages ​​is for disclosing the best mode of the present invention.

[0130] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0131] Similarly, it should be understood that in order to streamline the present disclosure and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the intention that the claimed invention requires more features than those explicitly recited in each claim. More specifically, as reflected in the claims below, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Therefore, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, with each claim itself serving as a separate embodiment of the present invention.

[0132] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0133] In addition, those skilled in the art will appreciate that, although some embodiments herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present invention and form different embodiments. For example, in the claims below, any one of the claimed embodiments may be used in any combination.

[0134] The various component embodiments of the present invention can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all functions of some or all components in a gateway, proxy server, or system according to an embodiment of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program implementing the present invention can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0135] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc., does not indicate any order. These words may be interpreted as names.

[0136] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0137] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for determining the oil-water two-phase permeability curve based on machine learning, It is characterized in that The method comprises: Acquire multiple sets of target parameters of developed reservoirs, and determine the first reference coefficients to be determined in the oil phase permeability curve model and the second reference coefficients to be determined in the water phase permeability curve model of the developed reservoirs; the target parameters include: rock parameters and fluid parameters; Constructing a sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined; Constructing a machine learning model, and training the machine learning model according to the sample data set to obtain a trained machine learning model; The trained and qualified machine learning model is used to predict the oil-water two-phase permeability curve of the target reservoir to obtain the oil-water two-phase permeability curve.

2. The method according to claim 1, It is characterized in that The method of determining a first reference coefficient to be determined in the oil phase relative permeability curve model of the developed reservoir and a second reference coefficient to be determined in the water phase relative permeability curve model comprises: For any set of target parameters, at different water saturations, a phase permeability curve simulation is performed based on the target parameters to obtain a corresponding oil phase phase permeability curve and a water phase phase permeability curve; Fitting a first function corresponding to the oil phase permeability curve and a second function corresponding to the water phase permeability curve; The first reference coefficient to be determined is extracted from the first function, and the second reference coefficient to be determined is extracted from the second function.

3. The method according to claim 2, It is characterized in that The first function is The second function is: The first reference coefficient to be determined includes a and b; the second reference coefficient to be determined includes c and d; the K ro is the oil phase relative permeability, the K o (S wi ) is the oil phase relative permeability corresponding to the irreducible water saturation, and the S wd is the dimensionless water saturation; rw is the relative permeability of water phase.

4. The method according to claim 1, It is characterized in that Before constructing the sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined, the method further includes: According to the formula Normalize the target parameters respectively to obtain the normalized target parameters If it is determined that the first reference coefficient to be determined and / or the second reference coefficient to be determined need to be processed with the same order of magnitude, the first reference coefficient to be determined and / or the second reference coefficient to be determined are reduced by a target multiple so that the values ​​of the first reference coefficient to be determined and the second reference coefficient to be determined are between [0, 1]; wherein, The z max is the maximum value of the target parameter before normalization, z min is the minimum value of the target parameter before normalization, and z is any target parameter before normalization.

5. The method according to claim 1, It is characterized in that The step of constructing a sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined includes: Taking each set of the target parameters as a reference data input set, and taking the corresponding first reference coefficient to be determined and the second reference coefficient to be determined as a reference data output set; The sample data set is constructed according to the reference data input set and the reference data output set.

6. The method according to claim 1, It is characterized in that The machine learning model is trained according to the sample data set to obtain a qualified machine learning model, including: Dividing the sample data set into a training set and a test set; The machine learning model is trained using the training set, and when the trained machine learning model is tested using the test set, if it is determined that the test accuracy meets the requirements, the corresponding machine learning model is output.

7. The method according to any one of claims 1 or 6, It is characterized in that The machine learning model is a deep neural network model, which includes an input layer, at least two hidden layers and an output layer; the deep neural network model is: The k is the output value of the kth layer of the deep neural network, and the output value of the output layer is the first undetermined coefficient and the second undetermined coefficient; x i is the input value of the i-th layer. When the i-th layer is the input layer, x i is the target parameter of the target reservoir; f o is the activation function of the output layer, f h is the activation function of the hidden layer, i is the number of the input layer neuron, j is the number of the hidden layer neuron, k is the number of the output layer neuron, N i is the number of neurons in the input layer, N h is the number of neurons in the output layer, W ij represents the connection weight between the i-th input layer neuron and the j-th hidden layer neuron, W jk represents the connection weight between the jth hidden layer neuron and the kth output layer neuron, (b h ) j represents the bias of the jth hidden layer neuron, (b o ) k Represents the bias of the kth output layer neuron.

8. The method according to claim 1, It is characterized in that The method of using the trained machine learning model to predict the oil-water two-phase permeability curve of the target reservoir includes: Using the target parameter of the target reservoir as an input value of the trained machine learning model, and using the trained machine learning model to output a first current undetermined coefficient and a second current undetermined coefficient; Determining an oil phase relative permeability curve model of the target oil reservoir according to the first currently undetermined coefficient, and determining a water phase relative permeability curve model of the target oil reservoir according to the second currently undetermined coefficient; The oil-water two-phase permeability curve of the target oil reservoir is determined according to the oil phase permeability curve model of the target oil reservoir and the water phase permeability curve model of the target oil reservoir.

9. A device for determining the oil-water two-phase permeability curve based on machine learning, It is characterized in that The device comprises: A determination unit is used to obtain multiple groups of target parameters of developed reservoirs, and determine the first reference coefficient to be determined in the oil phase permeability curve model and the second reference coefficient to be determined in the water phase permeability curve model of the developed reservoirs; the target parameters include: rock parameters and fluid parameters; A construction unit, configured to construct a sample data set according to the target parameter and the first reference coefficient to be determined and the second reference coefficient to be determined; A training unit, used to construct a machine learning model, train the machine learning model according to the sample data set, and obtain a qualified machine learning model; The prediction unit is used to predict the oil-water two-phase permeability curve of the target reservoir by using the trained machine learning model to obtain the oil-water two-phase permeability curve.

10. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images