A method and device for training an oil and gas reservoir quality classification model

Abstract

The application discloses a training method and device of an oil and gas reservoir quality classification model, which can be applied to the technical field of oil and natural gas engineering. The oil and gas reservoir quality classification model can be trained from four aspects of reservoir rock types, physical properties, longitudinal distribution and fluid types, so that the oil and gas reservoir quality classification model can comprehensively evaluate the classification of the reservoir. By comprehensively utilizing deep learning technology and geological and geophysical knowledge, the reservoir quality classification can be more comprehensive and accurate, scientific basis and technical support can be provided for oil and gas exploration and development, and the development scheme of the oil and gas reservoir is optimized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of oil and gas engineering, in particular to a training method and device of an oil and gas reservoir quality classification model. BACKGROUND

[0002] Reservoir quality classification refers to a process of dividing reservoirs into different quality grades through comprehensive analysis of physical properties (such as porosity, permeability, oil saturation, etc.) of the reservoirs. Reservoir quality classification is a key link in oil and gas exploration and development, which directly affects the development strategy, well pattern design and economic benefit of oil and gas fields.

[0003] At present, there is a lack of effective method for classifying reservoir quality, and therefore how to classify reservoir quality is a technical problem to be solved by those skilled in the art. SUMMARY

[0004] The present application provides a training method of an oil and gas reservoir quality classification model to classify reservoir quality, and also provides a training device of an oil and gas reservoir quality classification model.

[0005] In a first aspect, the present application provides a training method of an oil and gas reservoir quality classification model, comprising:

[0006] classifying reservoir quality of a target interval according to reservoir physical property parameters of the target interval and encoding classification labels obtained by classification to obtain label data of reservoir classes, the reservoir physical property parameters including wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness and fluid type;

[0007] obtaining seismic interval data of the target interval, and extracting the seismic interval data to obtain seismic trace data beside known wells;

[0008] counting value range distribution of different classes of reservoir physical property parameters of the known wells, generating a reservoir geological model according to the value range distribution of the reservoir physical property parameters and forward generating corresponding seismic data to obtain a reservoir template data set;

[0009] training a pre-constructed oil and gas reservoir quality classification model according to the label data, the seismic trace data and the reservoir template data set.

[0010] Optionally, the model includes A convolutional blocks, B ResNet blocks, a fully connected layer and a softmax layer, and the method further comprises:

[0011] obtaining non-neighboring well seismic data;

[0012] inputting the non-neighboring well seismic data into the trained oil and gas reservoir quality classification model for prediction.

[0013] Optionally, the reservoir quality of the target layer section is classified according to the reservoir property parameters of the target layer section, and a classification label obtained by classification is encoded to obtain label data of a reservoir category, including:

[0014] The reservoir quality of the target layer section is classified by using the reservoir property parameters and a reservoir quality classification standard of the target layer section, to obtain a reservoir quality classification label;

[0015] The reservoir quality classification label is one-hot encoded to obtain label data of a reservoir category, and a data dimension of the label data is (N, M), N is a number of wells, and M is a reservoir category.

[0016] Optionally, the seismic layer section data of the target layer section is obtained, and seismic trace data beside a known well is obtained by extracting the seismic layer section data, including:

[0017] The top and bottom seismic horizons of the target layer section are used as restrictions to generate seismic layer section data S1 of the target layer section;

[0018] The S1 is subjected to feature extraction to obtain seismic feature data S2 of the target layer section;

[0019] The seismic trace data beside the known well S3 is extracted from the S2 according to coordinates of the known well point.

[0020] Optionally, the reservoir property parameter value range distribution includes value range distributions of wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, longitudinal reservoir thickness, and fluid type, a reservoir geological model is generated according to the reservoir property parameter value range distribution, and corresponding seismic data is generated by forward modeling to obtain reservoir template data, including:

[0021] A normal distribution function of each reservoir property parameter is determined according to a mean value and a variance of the reservoir property parameter value range distribution;

[0022] A plurality of groups of reservoir property parameters are generated according to the normal distribution function of each reservoir property parameter to obtain the reservoir template data.

[0023] Optionally, the method further includes:

[0024] The wave impedance parameter of the reservoir template is input into a convolution model to obtain seismic waveforms corresponding to the plurality of groups of reservoir templates;

[0025] The pre-constructed oil and gas reservoir quality classification model is trained according to the label data, the seismic trace data, and the reservoir template data set, including:

[0026] training a pre-constructed oil and gas reservoir quality classification model according to the label data, the seismic trace data, the reservoir template data set and the seismic waveform

[0027] In a second aspect, the application further provides a device for training an oil and gas reservoir quality classification model, the device comprising:

[0028] a coding unit configured to classify reservoir quality of a target interval according to reservoir physical property parameters of the target interval, and encode classification labels obtained by the classification to obtain label data of reservoir categories, the reservoir physical property parameters comprising wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness and fluid type;

[0029] an extraction unit configured to obtain seismic interval data of the target interval, and extract the seismic interval data to obtain seismic trace data of known wells;

[0030] a generation unit configured to count value domain distribution of different categories of reservoir physical property parameters of the known wells, generate a reservoir geological model according to the value domain distribution of the reservoir physical property parameters, and forward generate corresponding seismic data to obtain a reservoir template data set;

[0031] a training unit configured to train a pre-constructed oil and gas reservoir quality classification model according to the label data, the seismic trace data and the reservoir template data set.

[0032] Optionally, the oil and gas reservoir quality classification model comprises A convolutional blocks, B ResNet blocks, a full connection layer and a softmax layer, and the device further comprises:

[0033] an acquisition unit configured to acquire non-well seismic data;

[0034] a prediction unit configured to input the non-well seismic data into the trained oil and gas reservoir quality classification model for prediction.

[0035] Optionally, the coding unit is specifically configured to:

[0036] classify reservoir quality of the target interval according to the reservoir physical property parameters and reservoir quality classification standards of the target interval to obtain reservoir quality classification labels;

[0037] one-hot encode the reservoir quality classification labels to obtain label data of reservoir categories, the label data having a data dimension of (N, M), N being a number of wells and M being a number of reservoir categories.

[0038] Optionally, the extraction unit is specifically configured to:

[0039] S1 is generated by using the top and bottom seismic horizons of the target layer as a limit;

[0040] S1 is subjected to feature extraction to obtain seismic feature data S2 of the target layer;

[0041] S3 is extracted from S2 according to the coordinates of the known well points

[0042] Optionally, the value range distribution of the reservoir physical property parameters includes value range distributions of wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness and fluid type, and the generating unit is specifically used for:

[0043] The normal distribution function of each reservoir physical property parameter is determined according to the mean and variance of the reservoir physical property parameter value range distribution;

[0044] A plurality of sets of reservoir physical property parameters are generated according to the normal distribution function of each reservoir physical property parameter, and reservoir template data is obtained.

[0045] Optionally, the device further includes:

[0046] The input unit is configured to input the wave impedance parameter of the reservoir template to the convolution model to obtain seismic waveforms corresponding to the plurality of sets of reservoir templates;

[0047] The training unit is specifically configured to:

[0048] The pre-constructed oil and gas reservoir quality classification model is trained according to the label data, the seismic trace data, the reservoir template data set and the seismic waveforms.

[0049] In a third aspect, an embodiment of the present application provides a device, the device including a memory and a processor, the memory being configured to store instructions or codes, and the processor being configured to execute the instructions or codes to enable the device to perform the method of the first aspect.

[0050] In a fourth aspect, an embodiment of the present application provides a computer storage medium, the computer storage medium storing codes, and when the codes are executed, a device executing the codes implements the method of the first aspect. BRIEF DESCRIPTION OF DRAWINGS

[0051] To make the technical solutions in the embodiments or the prior art clearer, the drawings needed in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0052] Figure 1 A flowchart of a training method of an oil and gas reservoir quality classification model provided by an embodiment of the present application is shown in FIG. 1.

[0053] Figure 2 A reservoir planar distribution map provided by an embodiment of the present application is shown in FIG. 2.

[0054] Figure 3 A structural schematic diagram of a specific embodiment of a training device of an oil and gas reservoir quality classification model provided by the present application is shown in FIG. 3. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0056] It should be noted that the training method and device of an oil and gas reservoir quality classification model provided by the present application are used in the application scenario of reservoir quality classification prediction. The above is only an example, and does not limit the application field of the method and device provided by the present application.

[0057] Reservoir quality classification refers to the process of dividing reservoirs into different quality grades through comprehensive analysis of the physical properties (such as porosity, permeability, oil saturation, etc.) of the reservoirs. Reservoir quality classification is a key link in oil and gas exploration and development, which directly affects the development strategy, well pattern design and economic benefit of oil and gas fields.

[0058] At present, there is a lack of effective method for classifying reservoir quality, and therefore how to classify reservoir quality is a technical problem to be solved by those skilled in the art.

[0059] Therefore, the application provides a training method of an oil and gas reservoir quality classification model, which comprises the following steps: classifying reservoir quality of a target interval according to reservoir physical parameters of the target interval and encoding classification labels obtained in the classification to obtain label data of reservoir categories, wherein the reservoir physical parameters comprise wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, longitudinal reservoir thickness and fluid type; obtaining seismic interval data of the target interval, extracting the seismic interval data to obtain seismic trace data beside known wells, counting reservoir physical parameter value range distribution of different categories of the known wells, generating a reservoir geological model according to the reservoir physical parameter value range distribution and forward generating corresponding seismic data to obtain a reservoir template data set, and training a pre-constructed oil and gas reservoir quality classification model according to the label data, the seismic trace data and the reservoir template data set. In the application, the oil and gas reservoir quality classification model can be trained from four aspects of reservoir rock type, physical property, longitudinal distribution and fluid type, so that the oil and gas reservoir quality classification model can comprehensively evaluate the classification of the reservoir, the reservoir quality classification can be more comprehensive and accurate by comprehensively utilizing deep learning technology and geological and geophysical knowledge, and scientific basis and technical support can be provided for oil and gas exploration and development, and an oil and gas reservoir development scheme can be optimized.

[0060] The method provided in the embodiments of the application can be executed by software on a computing device. The computing device can be, for example, a mobile phone, a tablet computer, a computer or the like.

[0061] In order for those skilled in the art to better understand the scheme of the application, the application will be further described in detail below in combination with the drawings and specific embodiments. The method provided in the embodiments of the application is taken as an example for description.

[0062] Figure 1 A flowchart of the training method of the oil and gas reservoir quality classification model provided in the embodiments of the application is shown in FIG. 1. The embodiment can be referred to as embodiment one, and as shown in the figure, the method comprises the following steps. Figure 1

[0063] S101: classifying reservoir quality of a target interval according to reservoir physical parameters of the target interval and encoding classification labels obtained in the classification to obtain label data of reservoir categories.

[0064] The computing device can obtain reservoir physical parameters of the target interval according to comprehensive interpretation of the known well curves (the reservoir physical parameters are parameters for describing physical properties of reservoir rocks, which directly reflect the quality and oil and gas storage capacity of the reservoir), and then can classify reservoir quality of the target interval according to the obtained reservoir physical parameters and encode classification labels obtained in the classification to obtain label data of reservoir categories. In the embodiment, the number of known wells can be 148. ​

[0065] In some possible implementation manners, the reservoir physical property parameters can include at least one of wave impedance (the wave impedance is the product of the density of rock and the seismic wave velocity when the seismic wave propagates in the rock medium), lithology type (the lithology type refers to the type of rock, such as sandstone, mudstone, carbonate rock, etc.), sedimentary facies type (the sedimentary facies type refers to the type of sedimentary environment, such as delta facies, river facies, lake facies, marine facies, etc.), porosity (the porosity refers to the percentage of the volume of pores in rock to the total volume of rock, reflecting the storage space of fluid in rock), permeability (the permeability refers to the ability of fluid flow in rock, reflecting the conductivity of fluid in rock), oil saturation (the oil saturation refers to the percentage of the volume of oil in rock to the pore volume, reflecting the content of oil in rock), vertical reservoir thickness (the vertical reservoir thickness refers to the thickness of the reservoir in the vertical direction, reflecting the scale of the reservoir), and fluid type (the fluid type refers to the type of fluid in the reservoir, such as oil, gas, and water). In the embodiments of the present application, all of the above are taken as examples for introduction.

[0066] The reservoir quality of the target layer section of each well can be classified by using the reservoir physical property parameters and the reservoir quality classification standard of the target layer section (the reservoir quality classification is taken as an example of three categories of type one reservoir, type two reservoir, and type three reservoir), to obtain a reservoir quality classification label, wherein the reservoir quality classification label is N, and the reservoir quality classification label is one-hot encoded to obtain label data of the reservoir category, the data dimension of the label data is (N, M), N is the number of wells, and M is the reservoir category. That is, the reservoir category label of the well point of the target layer section is processed by using the one-hot encoding method, and the string category (type one reservoir, type two reservoir, and type three reservoir) is converted into an integer format: {“type one reservoir”:[1,0,0], “type two reservoir”:[0,1,0], “type two reservoir”:[0,0,1]}, and the output format is (N, 3).

[0067] S102: Obtain seismic layer section data of a target layer section, and extract the seismic layer section data to obtain seismic trace data beside a known well.

[0068] The computing device can obtain seismic layer section data of a target layer section, and then extract the seismic layer section data to obtain seismic trace data beside a known well.

[0069] In some possible implementation manners, the computing device can generate target layer section seismic section data S1 using the top and bottom seismic horizons of the target layer section as constraints, perform feature extraction on S1 by using a wavelet transform method to obtain target layer section seismic feature data S2, and extract known well-side seismic trace data S3 from S2 according to the coordinates of the known well points, where the data dimension of S1 can be (T, H, W), the data dimension of S2 can be (F, T, H, W), and the data dimension of S3 can be (F, T, M).

[0070] For example, the data dimension of S1 can be (40, 400, 600), the data dimension of S2 can be (102, 40, 400, 600), and the data dimension of S3 can be (120, 40, 148).

[0071] For example, formula (1) can be used to process each seismic trace of data S1 respectively. After the processing of each seismic trace is completed in a loop, the target layer section seismic data features are obtained, and the output data S2 in the CWT(a, b) format is (F, T, H, W), which is named Seismic_F.

[0072] (1)

[0073] Wherein:

[0074] b: used for positioning time. a: used for positioning frequency. f(t): represents the original seismic waveform. represents a wavelet basis. In the wavelet basis, the variable a represents a scale corresponding to a frequency. The variable b represents a time shift.

[0075] S103: Statistically analyze the value range distribution of the physical parameters of different types of reservoirs of the known wells, generate a reservoir geological model according to the value range distribution of the physical parameters of the reservoirs, and forwardly generate corresponding seismic data to obtain a reservoir template data set.

[0076] The computing device can statistically analyze the value range distribution of the physical parameters of different types of reservoirs of the known wells, generate a reservoir geological model according to the value range distribution of the physical parameters of the reservoirs, and forwardly generate corresponding seismic data to obtain a reservoir template data set.

[0077] For example, according to the known 148 wells, the value range distribution of 8 types of reservoir physical parameters, such as wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness, and fluid type, of 3 types of reservoirs can be statistically analyzed. Then, according to the statistical results of the 8 types of parameters of the three types of reservoirs, 8 types of reservoir physical parameter values of the three types of reservoirs can be randomly generated, different types of reservoir geological models can be made, corresponding seismic data can be forwardly generated, and finally a reservoir template data set can be generated.

[0078] The wave impedance mean value of one of the reservoirs can be 6000-7000 (kg / m^3·m / s); the lithology type can mainly be medium-fine grained sandstone, and the porosity is relatively developed; the sedimentary facies type can mainly be distributary channel, mouth bar, etc.; the porosity mean value can be 25-30%; the permeability mean value can be 1000-3000 mD; the oil saturation mean value can be 60-80%, and the variance is small; the longitudinal reservoir thickness mean value can be 15-30 meters; and the fluid type can mainly be light oil.

[0079] The wave impedance mean value of the second type of reservoir can be 5000-6000 (kg / m^3·m / s); the lithology type can mainly be fine grained sandstone and siltstone; the sedimentary facies type can mainly be interdistributary bay, frontal sand dam, etc.; the porosity mean value can be 15-25%; the permeability mean value can be 100-1000 mD; the oil saturation mean value can be 40-60%; the longitudinal reservoir thickness mean value can be 10-20 meters; and the fluid type can mainly be medium oil.

[0080] The wave impedance mean value of the third type of reservoir can be 4000-5000 (kg / m^3·m / s); the lithology type can mainly be siltstone and argillaceous sandstone; the sedimentary facies type can mainly be flood plain, lacustrine facies, etc.; the porosity mean value can be 5-15%; the permeability mean value can be 10-100 mD; the oil saturation mean value can be 20-40%; the longitudinal reservoir thickness mean value can be 5-10 meters; and the fluid type can mainly be heavy oil or aquifer.

[0081] Specifically, the normal distribution function of each reservoir physical property parameter can be determined according to the mean value and variance of the value range distribution of the physical property parameter, and then a plurality of groups (here, 10000 groups are taken as an example) of reservoir physical property parameters are generated according to the normal distribution function of each reservoir physical property parameter, to obtain the reservoir template data.

[0082] The normal distribution function of each reservoir physical property parameter according to the mean value and variance of the reservoir physical property parameter can be specifically as shown in formula (2):

[0083] (2)

[0084] wherein, : represents the mean value of different reservoir properties. : represents the variance of different reservoir properties. x: represents different reservoir property parameters, : normalization constant, ensuring the integral (i.e., total probability) of the entire distribution is 1, : exponential term, determining the shape of the distribution. It represents the degree of deviation of x from the mean value , which is scaled by the standard deviation .

[0085] By creating a reservoir geological model and generating a reservoir template dataset, we provide additional data with prior knowledge for model training. This data can help the model learn about seismic forward modeling and reservoir classification standards, enhancing its understanding and generalization capabilities for reservoir quality classification.

[0086] In some possible implementations, the computing device may also use the wave impedance parameters of the reservoir template and utilize the convolution model, i.e., formula (3), to obtain multiple groups (10,000 groups in this embodiment as an example) of seismic waveforms S4 corresponding to the reservoir templates.

[0087] (3)

[0088] in: : Reservoir templates generate seismic waveforms. : Ricker wavelet. : Reservoir template parameter wave impedance difference to obtain reflection coefficient.

[0089] S104: Training a pre-built oil and gas reservoir quality classification model according to the label data, the seismic trace data, and the reservoir template data set.

[0090] First, let's introduce the structure of the pre-built oil and gas reservoir quality classification model, which can include two convolutional blocks and three ResNet blocks. The first and second convolutional blocks of the reservoir quality classification model can be composed of a convolutional layer with a 1*3 convolution kernel, a BatchNormal layer, and a ReLU activation layer, respectively. The three ResNet blocks of the reservoir quality classification model can be composed of two convolutional layers with a 1*1 convolution kernel, a convolutional layer with a 1*3 convolution kernel, a ReLU activation function, and a skip connection structure, respectively, as shown in Formula (4):

[0091] (4)

[0092] in: : Standard ReLU activation function. BN: represents the BatchNormal normalization layer.

[0093] : Represents the parameters to be learned for the three convolutional layers. x: Represents the input seismic data features. y: Represents the output seismic data features.

[0094] The last layer of the reservoir quality classification model uses a fully connected layer, and the final activation function used is softmax, as shown in formula (5). The cross entropy loss function can be used as the optimization objective function of the reservoir quality classification model, as shown in formula (6):

[0095] (5)

[0096] denotes the i-th component in the input vector. denotes the probability corresponding to each class.

[0097] (6)

[0098] where: denotes the indicator function (0 or 1); takes 1 if the true class of sample point i is equal to c, otherwise takes 0. denotes the predicted probability that the observed sample point i belongs to class c. M denotes the number of classes. N denotes the number of samples.

[0099] In the embodiments of the present application, the oil and gas reservoir quality classification model can be trained according to the label data, the seismic trace data and the reservoir template data set.

[0100] Specifically, training the oil and gas reservoir quality classification model can include two steps, first, the oil and gas reservoir quality classification model (hereinafter referred to as the model) can be trained using the reservoir template data set. This step of training the model mainly adds prior knowledge about seismic forward and reservoir division criteria to the model. The hyperparameters used when training the reservoir quality classification model can be learning rate set to 0.001, epoch set to 1000, optimization function set to Adam optimization function, and batch set to 64.

[0101] In some possible implementations, in the first step of the training process, a plurality of sets (10000 sets in this embodiment) of seismic waveforms S4 data corresponding to the reservoir templates can also be added to the model, that is, adding this part of data to train the model can enable the model to more accurately identify and classify reservoir quality and enhance the generalization ability of the model.

[0102] The second step can use the known seismic trace data beside the well to train the oil and gas reservoir quality classification model. This step is to prepare the downstream task model according to the specific experimental work area, and more specifically, to prepare the downstream task model. The hyperparameters used when training the downstream task M1 layer reservoir classification model can be learning rate set to 0.00001, epoch set to 2000, optimization function set to Adam optimization function, and batch set to 64. In this way, the training of the oil and gas reservoir quality classification model can be realized.

[0103] In some possible implementation manners, non-well seismic data can be acquired. The non-well seismic data refer to seismic data acquired in an area without drilled wells, and the data can be used to predict and evaluate the quality of a reservoir. The non-well seismic data are input into the trained oil and gas reservoir quality classification model, and the reservoir category prediction of all seismic traces is completed in a loop. Finally, the reservoir categories corresponding to the seismic traces are combined according to the positions of the seismic traces, to form a reservoir planar distribution map of the M1 layer section, as shown in FIG. 8. Green represents a first type of reservoir, gray represents a second type of reservoir, and blue represents a third type of reservoir. Figure 3

[0104] In this embodiment, the reservoir quality of a target layer section can be classified according to reservoir physical parameters of the target layer section, and a label data of a reservoir category is obtained by encoding a classification label obtained in the classification. The reservoir physical parameters include wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness, and fluid type. Seismic layer section data of the target layer section is acquired, and seismic trace data beside known wells is obtained by extraction. The value domain distribution of reservoir physical parameters of different categories of the known wells is counted, a reservoir geological model is generated according to the value domain distribution of the reservoir physical parameters, and corresponding seismic data is generated by forward modeling, to obtain a reservoir template data set. The pre-constructed oil and gas reservoir quality classification model is trained according to the label data, the seismic trace data, and the reservoir template data set. In this application, the oil and gas reservoir quality classification model can be trained from four aspects of reservoir rock type, physical property, vertical distribution, and fluid type, so that the oil and gas reservoir quality classification model can more comprehensively evaluate the classification of the reservoir. By comprehensively utilizing deep learning technology and geological and geophysical knowledge, the reservoir quality classification can be more comprehensive and accurate, to provide a scientific basis and technical support for oil and gas exploration and development, and to optimize an oil and gas reservoir development scheme.

[0105] The above is some specific implementation manners of the training method of the oil and gas reservoir quality classification model provided in the embodiments of this application. Based on this, the application further provides a corresponding device. The device provided in the embodiments of this application will be introduced from the perspective of functional modularization. The device can be mutually referred to the training method of the oil and gas reservoir quality classification model described above.

[0106] Figure 3 The structure block diagram of the training device of the oil and gas reservoir quality classification model provided in the embodiments of this application is called specific implementation manner two, and is referred to FIG. 9. Figure 3 The device can include:

[0107] ​The encoding unit 300 is configured to classify reservoir quality of the target layer section according to reservoir physical property parameters of the target layer section, and encode the classification label obtained by the classification to obtain label data of the reservoir category, wherein the reservoir physical property parameters include wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness and fluid type.

[0108] The extraction unit 310 is configured to acquire seismic section data of the target layer section, and extract the seismic section data to obtain seismic trace data beside the known well;

[0109] The generation unit 320 is configured to count value domain distribution of different categories of reservoir physical property parameters of the known well, generate a reservoir geological model according to the value domain distribution of the reservoir physical property parameters, and forward generate corresponding seismic data to obtain a reservoir template data set;

[0110] The training unit 330 is configured to train a pre-constructed oil and gas reservoir quality classification model according to the label data, the seismic trace data and the reservoir template data set.

[0111] Optionally, the oil and gas reservoir quality classification model includes A convolution blocks, B ResNet blocks, a full connection layer and a softmax layer, and the apparatus further includes:

[0112] The acquisition unit is configured to acquire non-well seismic data;

[0113] The prediction unit is configured to input the non-well seismic data into the trained oil and gas reservoir quality classification model for prediction.

[0114] Optionally, the encoding unit is specifically configured to:

[0115] classify the reservoir quality of the target layer section according to the reservoir physical property parameters and a reservoir quality classification standard of the target layer section to obtain a reservoir quality classification label;

[0116] one-hot encode the reservoir quality classification label to obtain label data of the reservoir category, wherein a data dimension of the label data is (N, M), N is a well number, and M is a reservoir category.

[0117] Optionally, the extraction unit is specifically configured to:

[0118] use top and bottom seismic horizons of the target layer section as a limit to generate seismic section data S1 of the target layer section;

[0119] perform feature extraction on S1 to acquire seismic feature data S2 of the target layer section;

[0120] extract seismic trace data S3 beside the known well from S2 according to coordinates of the known well point

[0121] Optionally, the reservoir property parameter value domain distribution includes value domain distributions of wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness and fluid type, and the generating unit is specifically used for:

[0122] determining a normal distribution function of each reservoir property parameter according to the mean and variance of the reservoir property parameter value domain distribution;

[0123] generating a plurality of groups of reservoir property parameters according to the normal distribution function of each reservoir property parameter, to obtain reservoir template data.

[0124] Optionally, the apparatus further includes:

[0125] The input unit is configured to input the wave impedance parameter of the reservoir template to the convolution model to obtain a plurality of groups of seismic waveforms corresponding to the reservoir template.

[0126] The training unit is specifically configured to:

[0127] train a pre-constructed oil and gas reservoir quality classification model according to the label data, the seismic trace data, the reservoir template data set and the seismic waveforms.

[0128] The embodiments of the present application further provide a corresponding device and a computer storage medium for implementing the schemes provided by the embodiments of the present application.

[0129] The device includes a memory and a processor, the memory is configured to store instructions or codes, and the processor is configured to execute the instructions or codes to enable the device to perform the method described in any of the embodiments of the present application.

[0130] The computer storage medium stores codes, and when the codes are executed, a device running the codes implements the method described in any of the embodiments of the present application.

[0131] The terms “first” and “second” in the names mentioned in the embodiments of the present application are only used for name identification, and do not represent the first and second in order.

[0132] Through the description of the above embodiments, it can be seen that those skilled in the art can clearly understand that all or part of the steps in the above-mentioned embodiment methods can be implemented by means of software plus a general hardware platform. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product. The computer software product can be stored in a storage medium, such as a read-only memory (ROM) / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network communication device such as a router) to execute the methods described in various embodiments or certain parts of the embodiments of the present application.

[0133] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For the relevant parts, refer to the partial description of the method embodiment. Some or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. Those of ordinary skill in the art can understand and implement it without paying any creative work.

[0134] The above description is merely an exemplary embodiment of the present application and is not intended to limit the scope of protection of the present application.

Claims

1. A training method for an oil and gas reservoir quality classification model, characterized in that: include: Classifying the reservoir quality of the target layer according to the reservoir physical property parameters of the target layer and encoding the classification labels obtained by classification to obtain label data of the reservoir category, wherein the reservoir physical property parameters include: wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness and fluid type; Using the top and bottom seismic horizons of the target interval as constraints, generating seismic interval data of the target interval; The seismic layer segment data of the target layer segment is subjected to feature extraction using the following formula to obtain the seismic characteristic data of the target layer segment: ； Among them, CWT represents continuous wavelet transform, a is used to locate frequency, b is used to locate time, F is the function space representation corresponding to the original earthquake waveform f(t), f(t) represents the original earthquake waveform, Represents the wavelet basis, the variable a in the wavelet basis represents the scale, and the translation b represents the time; Extracting known wellside seismic trace data from the seismic characteristic data of the target layer according to the coordinates of the known well point; Statistically analyzing the distribution of reservoir physical property parameters of different categories in known wells, generating a reservoir geological model based on the distribution of reservoir physical property parameters, and forward modeling corresponding seismic data to obtain a reservoir template data set; Input the wave impedance parameters of the reservoir template into the convolution model to obtain multiple sets of seismic waveforms corresponding to the reservoir templates; A pre-built oil and gas reservoir quality classification model is trained based on the label data, the seismic trace data, the reservoir template data set, and the seismic waveform.

2. The method according to claim 1, characterized in that The oil and gas reservoir quality classification model includes multiple convolution blocks, multiple ResNet blocks, a fully connected layer, and a softmax layer. The method further includes: Acquire non-adjacent well seismic data; The non-adjacent well seismic data is input into a trained oil and gas reservoir quality classification model for prediction.

3. The method according to claim 1, characterized in that The step of classifying the reservoir quality of the target layer according to the reservoir physical property parameters of the target layer and encoding the classification labels obtained by classification to obtain label data of the reservoir category includes: Classifying the reservoir quality of the target layer section using the reservoir physical property parameters and the reservoir quality classification standard of the target layer section to obtain a reservoir quality classification label; One-hot encoding is performed on the reservoir quality classification label to obtain label data of the reservoir category. The data dimension of the label data is (N, M), where N is the number of wells and M is the reservoir category.

4. The method according to claim 1, wherein The reservoir physical property parameter value range distribution includes: wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness and fluid type value range distribution. Based on the reservoir physical property parameter value range distribution, a reservoir geological model is generated and corresponding seismic data is forward-modeled to obtain reservoir template data, including: Determining a normal distribution function for each reservoir physical property parameter based on the mean and variance of the reservoir physical property parameter range distribution; According to the normal distribution function of each reservoir physical property parameter, multiple groups of reservoir physical property parameters are generated to obtain reservoir template data.

5. A training device for an oil and gas reservoir quality classification model, characterized in that: include: an encoding unit, configured to classify the reservoir quality of the target layer according to the reservoir physical property parameters of the target layer and encode the classification labels obtained by classification to obtain label data of the reservoir category, wherein the reservoir physical property parameters include: wave impedance, lithology type, sedimentary facies type, porosity, permeability, oil saturation, vertical reservoir thickness and fluid type; The extraction unit is used to use the top and bottom seismic horizons of the target layer segment as a restriction to generate seismic layer segment data of the target layer segment; and extract features of the seismic layer segment data of the target layer segment using the following formula to obtain seismic feature data of the target layer segment: ； Among them, CWT represents continuous wavelet transform, a is used to locate frequency, b is used to locate time, F is the function space representation corresponding to the original earthquake waveform f(t), f(t) represents the original earthquake waveform, represents a wavelet basis, wherein the variable a in the wavelet basis represents scale, and the translation b represents time; according to the coordinates of the known well point, the seismic trace data near the known well is extracted from the seismic characteristic data of the target layer; A generation unit is used to collect statistics on the range distribution of reservoir physical property parameters of different categories in known wells, generate a reservoir geological model based on the range distribution of reservoir physical property parameters, and forward-model corresponding seismic data to obtain a reservoir template data set; An input unit, used to input the wave impedance parameters of the reservoir template into the convolution model to obtain seismic waveforms corresponding to multiple groups of reservoir templates; A training unit is used to train a pre-built oil and gas reservoir quality classification model based on the label data, the seismic trace data, the reservoir template data set and the seismic waveform.

6. The device according to claim 5, characterized in that The oil and gas reservoir quality classification model includes multiple convolution blocks, multiple ResNet blocks, a fully connected layer and a softmax layer. The device also includes: An acquisition unit, used for acquiring seismic data of non-adjacent wells; The prediction unit is used to input the non-adjacent well seismic data into a trained oil and gas reservoir quality classification model for prediction.

7. The device according to claim 5, characterized in that The encoding unit is specifically used to: Classifying the reservoir quality of the target layer section using the reservoir physical property parameters and the reservoir quality classification standard of the target layer section to obtain a reservoir quality classification label; One-hot encoding is performed on the reservoir quality classification label to obtain label data of the reservoir category. The data dimension of the label data is (N, M), where N is the number of wells and M is the reservoir category.

Images