A sedimentary microfacies identification method, system, device and computer storage medium

Abstract

The present application discloses a sedimentary microfacies identification method, system, equipment and computer storage medium, which relates to the field of oil development technology, and includes obtaining target logging data of a well point to be identified; generating target logging phase element values ​​of the well point to be identified based on the target logging data of the well point to be identified; dividing the well point to be identified into sample well points and test well points; determining the sedimentary microfacies type of the sample well point based on the target logging phase element values ​​of the sample well points; using the target logging data of the sample well point and the target logging phase element values ​​of the sample well point as inputs, and the sedimentary microfacies type of the sample well point as outputs, training a deep learning network, and obtaining a trained sedimentary microfacies identification model; inputting the target logging data of the test well point and the target logging phase element values ​​of the test well point into the sedimentary microfacies identification model, and obtaining the sedimentary microfacies type of the test well point. The present application uses the sedimentary microfacies identification model to identify the sedimentary microfacies type, thereby improving efficiency and accuracy.

Description

Technical Field

[0001] The present application relates to the technical field of petroleum development, and more specifically, to a sedimentary microfacies identification method, system, equipment and computer storage medium. Background Art

[0002] In oilfield development, the overall sedimentary microfacies distribution of the study area can be inferred from the single-well sedimentary microfacies of each oil well. By studying the sedimentary microfacies distribution of the study area, the distribution law of oil and gas in the formation can be analyzed, which is helpful for reasonable reservoir evaluation and development planning. In the identification of sedimentary microfacies, the single-well sedimentary microfacies of several standard wells are often manually determined first, and then the single-well sedimentary microfacies of the remaining wells in the study area are identified one by one using the logging curve pattern of the standard wells.

[0003] However, the identification process of sedimentary microfacies involves repeated comparison work, which is difficult to operate, inefficient, and prone to comparison errors, affecting the accuracy of identification of sedimentary microfacies.

[0004] In summary, how to accurately and efficiently identify sedimentary microfacies is an urgent problem to be solved by technicians in this field. Summary of the invention

[0005] The purpose of the present application is to provide a sedimentary microphase identification method, which can solve the technical problem of how to accurately and efficiently identify sedimentary microphases to a certain extent. The present application also provides a sedimentary microphase identification system, an electronic device and a computer-readable storage medium.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] A sedimentary microfacies identification method, comprising:

[0008] Acquire target logging data of the well point to be identified;

[0009] Generating a target logging phase element value of the well point to be identified according to the target logging data of the well point to be identified;

[0010] Dividing the well points to be identified into sample well points and test well points;

[0011] Determining the sedimentary microfacies type of the sample well point according to the target logging facies element value of the sample well point;

[0012] Taking the target logging data of the sample well point and the target logging phase element value of the sample well point as input, taking the sedimentary microfacies type of the sample well point as output, training the deep learning network to obtain a trained sedimentary microfacies identification model;

[0013] Inputting the target logging data of the test well point and the target logging phase element value of the test well point into the sedimentary microfacies identification model to obtain the sedimentary microfacies type of the test well point;

[0014] Among them, the target logging data include natural potential, natural gamma, micro gradient and micro electrical rate; the target logging phase element values ​​include logging curve amplitude, logging curve shape, thickness and curve smoothness.

[0015] Preferably, the target logging data of the sample well point and the target logging phase element value of the sample well point are used as input, and the sedimentary microfacies type of the sample well point is used as output to train the deep learning network to obtain a trained sedimentary microfacies identification model, including:

[0016] Determining the number of said sample well points for each sedimentary microfacies type;

[0017] Determine the weight of each sedimentary microfacies type based on the inverse ratio of said quantities;

[0018] Determine the initial deep learning network;

[0019] Inputting the target logging data of the sample well point and the target logging phase element value of the sample well point into the deep learning network;

[0020] Obtaining the predicted probability of the sedimentary microfacies type of the sample well point by the deep learning network;

[0021] Generate a loss value of the deep learning network based on the weight, the prediction probability, and the sedimentary microfacies type of the sample well point;

[0022] The deep learning network is updated based on the loss value to obtain a trained sedimentary microfacies identification model.

[0023] Preferably, the sedimentary microfacies identification model comprises a first same convolution block, a second same convolution block, a first downsampling layer, a third same convolution block, a second downsampling layer, a convolution layer, a connection layer and four fully connected layers connected in sequence; and the same convolution block comprises a same convolution layer, a normalization layer and an activation layer;

[0024] The input of the first convolutional block includes target logging data, and the input of the connection layer includes the output of the convolutional layer and the target logging phase element value.

[0025] Preferably, generating the logging curve amplitude of the well point to be identified according to the target logging data of the well point to be identified includes:

[0026] Get the set data interval;

[0027] Counting the proportion of the target logging data of the well point to be identified in each of the data intervals;

[0028] According to the size of the proportion, the target logging data of the well point to be identified are sorted to obtain a sorting result;

[0029] Filter out the first type of target logging data whose proportion is the target value from the sorting results;

[0030] The logging curve amplitude of the well point to be identified is generated according to the first type of target logging data.

[0031] Preferably, generating the logging curve form of the well point to be identified according to the target logging data of the well point to be identified includes:

[0032] Converting the target logging data of each of the well points to be identified into a target feature vector;

[0033] Generate the central value of each type of target logging data;

[0034] generating a center vector of target logging data according to the center value;

[0035] Generate a Mahalanobis distance between a target feature vector of each well point to be identified and the center vector;

[0036] The shape of the well logging curve of the well point to be identified is determined according to the Mahalanobis distance of the well point to be identified.

[0037] Preferably, generating the curve smoothness of the well point to be identified according to the target logging data of the well point to be identified includes:

[0038] Performing discrete Fourier transform on the target logging data of the well point to be identified to obtain initial frequency domain data;

[0039] Determining a frequency threshold according to the initial frequency domain data;

[0040] In the initial frequency domain data, frequency domain data having a frequency higher than the frequency threshold is selected as target frequency domain data;

[0041] Performing an inverse Fourier transform on the target frequency domain data to obtain a second type of target logging data;

[0042] generating a toothing amplitude of the second type of target logging data;

[0043] The smoothness of the curve of the well point to be identified is determined based on the toothing amplitude.

[0044] A sedimentary microfacies identification system, comprising:

[0045] A first acquisition module is used to acquire target logging data of a well point to be identified;

[0046] A first generating module, used for generating a target logging phase element value of the well point to be identified according to the target logging data of the well point to be identified;

[0047] A first division module is used to divide the well points to be identified into sample well points and test well points;

[0048] A first determination module is used to determine the sedimentary microfacies type of the sample well point according to the target logging facies element value of the sample well point;

[0049] The first training module is used to take the target logging data of the sample well point and the target logging phase element value of the sample well point as input, and take the sedimentary microfacies type of the sample well point as output, to train the deep learning network and obtain a trained sedimentary microfacies recognition model;

[0050] A first identification module is used to input the target logging data of the test well point and the target logging phase element value of the test well point into the sedimentary microfacies identification model to obtain the sedimentary microfacies type of the test well point;

[0051] Among them, the target logging data include natural potential, natural gamma, micro gradient and micro electrical rate; the target logging phase element values ​​include logging curve amplitude, logging curve shape, thickness and curve smoothness.

[0052] Preferably, the first training module includes:

[0053] A first determination unit is used to determine the number of the sample well points of each type of sedimentary microfacies;

[0054] A second determining unit is used to determine the weight of each type of sedimentary microfacies according to the inverse ratio of the quantity;

[0055] A third determining unit, used to determine an initial deep learning network;

[0056] A first input unit, used to input the target logging data of the sample well point and the target logging phase element value of the sample well point into the deep learning network;

[0057] A first acquisition unit is used to obtain the prediction probability of the sedimentary microfacies type of the sample well point by the deep learning network;

[0058] A first generating unit, configured to generate a loss value of the deep learning network based on the weight, the predicted probability, and the sedimentary microfacies type of the sample well point;

[0059] The first updating unit is used to update the deep learning network based on the loss value to obtain a trained sedimentary microfacies identification model.

[0060] An electronic device, comprising:

[0061] Memory for storing computer programs;

[0062] A processor is used to implement the steps of any of the above-mentioned sedimentary microfacies identification methods when executing the computer program.

[0063] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any of the above-mentioned sedimentary microfacies identification methods are implemented.

[0064] The present application provides a sedimentary microfacies identification method, which comprises the following steps: obtaining target logging data of a well point to be identified; generating target logging facies element values ​​of the well point to be identified based on the target logging data of the well point to be identified; dividing the well point to be identified into sample well points and test well points; determining the sedimentary microfacies type of the sample well point based on the target logging facies element values ​​of the sample well points; training a deep learning network using the target logging data of the sample well point and the target logging facies element values ​​of the sample well point as inputs and the sedimentary microfacies type of the sample well point as outputs to obtain a trained sedimentary microfacies identification model; inputting the target logging data of the test well point and the target logging facies element values ​​of the test well point into the sedimentary microfacies identification model to obtain the sedimentary microfacies type of the test well point. The present application divides the well points to be identified into sample well points and test well points, and obtains a sedimentary microfacies identification model through the target logging data, target logging phase element value and sedimentary microfacies type training of the sample well points, so that the sedimentary microfacies type of the test well points can be identified with the help of the sedimentary microfacies identification model, without manual operation, with a high degree of automation, improving the identification efficiency of sedimentary microfacies types, avoiding operational errors caused by manual identification, and improving the identification accuracy of sedimentary microfacies types. The sedimentary microfacies identification system, electronic device and computer-readable storage medium provided by the present application also solve the corresponding technical problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0066] Figure 1 A flow chart of a sedimentary microfacies identification method provided in an embodiment of the present application;

[0067] Figure 2 This is a schematic diagram of the data processing for this application;

[0068] Figure 3 This is the framework diagram of the sedimentary microfacies identification model;

[0069] Figure 4 A structural schematic diagram of a sedimentary microfacies identification system provided in an embodiment of the present application;

[0070] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0071] Figure 6 Another structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0072] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0073] See also Figure 1 and Figure 2 , Figure 1 A flow chart of a sedimentary microfacies identification method provided in an embodiment of the present application, Figure 2 This is a schematic diagram of the data processing for this application.

[0074] A sedimentary microfacies identification method provided in an embodiment of the present application may include the following steps:

[0075] Step S101: Acquire target logging data of the well point to be identified.

[0076] Step S102: Generate target logging phase element values ​​of the well point to be identified based on the target logging data of the well point to be identified.

[0077] In practical applications, the target logging data of the well point to be identified can be obtained first. The well point to be identified refers to the well point that needs to be identified for sedimentary microfacies type. The well point to be identified can be an oil field well and other well points. The logging data refers to the data used to reflect the lithology and sedimentary characteristics. In addition, the formation depth data and stratification data of the well point to be identified can be obtained for further analysis of the well point to be identified. And it is necessary to generate the target logging phase element value of the well point to be identified based on the target logging data of the well point to be identified, so that the sedimentary microfacies type of the well point can be analyzed based on the logging phase element value of the well point. The sedimentary microfacies type of the well point can be river channel microfacies, overflow sand microfacies, abandoned river channel microfacies, floodplain microfacies, etc.

[0078] In practical applications, the types of logging data and logging phase element values ​​can be flexibly determined according to the application scenarios. For example, in the process of obtaining target logging data of the well point to be identified, the target logging data of the well point to be identified can be obtained, and the target logging data include natural potential, natural gamma, micro-gradient and micro-electric rate, etc.; in the process of generating the target logging phase element value of the well point to be identified based on the target logging data of the well point to be identified, the target logging phase element value of the well point to be identified can be generated based on the target logging data of the well point to be identified, and the target logging phase element value includes logging curve amplitude, logging curve shape, thickness and curve smoothness, etc.

[0079] In specific application scenarios, considering that the amplitude of the logging curve only focuses on sandstone with high amplitude characteristics, and the logging data includes low-amplitude mudstone and high-amplitude sandstone, the threshold for dividing low amplitude and high amplitude can be determined according to actual needs. If all the logging data in the small layer are used for calculation, the amplitude of the target area will be averaged. In order to avoid operational situations, the calculation target should be concentrated on the high-amplitude sandstone, that is, in the process of generating the logging curve amplitude of the well point to be identified according to the target logging data of the well point to be identified, the set data interval can be obtained; the proportion of the target logging data of the well point to be identified in each data interval is counted; according to the size of the proportion, the target logging data of the well point to be identified is sorted to obtain the sorting result; the first type of target logging data with a proportion of the target value is selected from the sorting result. Assuming the target value is p, p is determined according to the data of the area where the well point is located. As long as the low-amplitude mudstone can be screened out, the first type of target logging data can be the target logging data. arrive The data between Indicates rounding up. Indicates rounding down, and n indicates the number of data values. According to the first type of target logging data, the logging curve amplitude of the well point to be identified is generated. For example, the root mean square amplitude of the first type of target logging data can be calculated as the logging curve amplitude of the well point to be identified. The formula for generating the root mean square amplitude can be: , A RMS represents the value of the RMS amplitude, x i Represents the i-th data in the target logging data.

[0080] In specific application scenarios, the morphology of a logging curve refers to the characteristic pattern of the appearance displayed by the logging curve in a layer. Common types of logging curve morphology include bell-shaped, box-shaped, funnel-shaped, etc., and the morphological characteristics of the logging curve can be quantitatively described and classified by calculating and counting the mean, variance, skewness, kurtosis and other data of the logging curve. In this process, since the Mahalanobis distance defines the category boundary in the ellipsoid region, takes into account the correlation between features and can process the covariance information of different category distributions, the Mahalanobis distance can be used to determine the morphology of the logging curve. That is, in the process of generating the logging curve morphology of the well point to be identified based on the target logging data of the well point to be identified, the target logging data of each well point to be identified can be converted into a target feature vector. For example, the target feature vector can be expressed as X=(Mean, Variance, Skewness, Kurtosis), where Mean represents the mean of the logging data of the well point, Variance represents the variance of the logging data, and Skewness represents the variance of the logging data. Data skewness, Kurtosis represents the kurtosis of logging data, and each type of logging data of each well point has these four types of values. Taking natural points as an example, the mean, variance, data skewness and kurtosis of natural points constitute a set of vectors of natural points; generate the central value of each type of target logging data, such as taking the average value of this type of logging data as the central value, etc. Still taking natural potential as an example, the average value of natural potential of all well points can be taken as the central value of natural potential, etc.; generate the central vector of target logging data according to the central value, that is, combine the central values ​​of each type of target logging data together as the central vector; generate the Mahalanobis distance between the target feature vector and the central vector of each well point to be identified, and the Mahalanobis distance generation formula can be: , represents the Mahalanobis distance value of the dth well point to be identified, represents the target feature vector of the dth well point to be identified, represents the central value of the b-th target logging data, ∑ b Represents the covariance matrix of category b to define the shape and range of the category, T represents the transpose; according to the Mahalanobis distance of the well point to be identified, the shape of the logging curve of the well point to be identified is determined, for example, by To determine the shape of the well logging curve, Indicates the type of logging curve shape. , They respectively represent the maximum and minimum values ​​of the logging curve morphology categories corresponding to each vector point.

[0081] In specific application scenarios, the smaller the toothing amplitude, the higher the smoothness of the logging curve; the larger the toothing amplitude, the lower the smoothness of the logging curve. Therefore, the smoothness of the logging curve can be determined based on the toothing amplitude. However, in some smooth logging curves, due to the sharp increase in the values ​​of adjacent points, the toothing amplitude value will give the illusion that the value is too large, resulting in errors in the determined curve smoothness. In order to avoid this situation, the logging curve can be scaled to reduce the degree of change of adjacent points. For example, a low-frequency filtering method can be used to filter out the low-frequency components of the logging curve to obtain the high-frequency components that require attention from the toothing amplitude. That is, in the process of generating the curve smoothness of the well point to be identified based on the target logging data of the well point to be identified, the target logging data of the well point to be identified can be subjected to discrete Fourier transform to obtain initial frequency domain data. The discrete Fourier transform can be expressed as , represents the frequency domain value corresponding to the frequency f, and N represents the logging data for discrete Fourier transform. The number of g represents the data subscript; the frequency threshold is determined according to the initial frequency domain data, for example, half of the highest frequency is determined as the frequency threshold; in the initial frequency domain data, the frequency domain data above the frequency threshold is selected as the target frequency domain data; the target frequency domain data is inversely transformed by Fourier transform to obtain the second type of target logging data. The inverse Fourier transform formula can be , represents the second type of target logging data, Representing target frequency domain data; generating the toothing amplitude of the second type of target logging data; determining the smoothness of the curve of the well point to be identified based on the toothing amplitude, for example, the toothing amplitude can be used as the smoothness of the curve, etc.

[0082] It should be noted that the generation process of the tooth amplitude can be flexibly determined according to the application scenario. For example, for each type of data, the gradient of each data can be calculated first to obtain the rate of change of the curve at each data. This gradient can be expressed by the formula Generated, Gradient[h] represents the gradient value of the hth data, x h+1 With x h-1 They represent the values ​​of the (h+1)th data and the (h-1)th data respectively; the calculated gradients are converted to absolute values ​​to ensure that all gradient values ​​are positive numbers; finally, the mean of the absolute valued gradients is calculated to obtain the toothing amplitude of the data.

[0083] It should also be noted that the target logging data and target logging phase element values ​​of the well point to be identified can both be standardized values, that is, after obtaining the initial logging data of the well point to be identified, the initial logging data can be standardized to obtain the target logging data, and after generating the initial logging phase element values ​​according to the target logging data, the initial logging phase element values ​​can be standardized to obtain the target logging phase element values. And the standardization method can be flexibly determined according to actual needs, for example, the formula for standardization can be , represents the standardized data, L represents the original data, and In addition, each type of logging data can be processed according to this scheme to generate the logging phase element value corresponding to this type of logging data, or all types of logging data can be processed through the matrix to directly generate the logging phase element value corresponding to each type of logging data.

[0084] Step S103: Divide the well points to be identified into sample well points and test well points.

[0085] Step S104: Determine the sedimentary microfacies type of the sample well point according to the target logging facies element value of the sample well point.

[0086] Step S105: Taking the target logging data of the sample well point and the target logging phase element value of the sample well point as input, and taking the sedimentary microfacies type of the sample well point as output, the deep learning network is trained to obtain a trained sedimentary microfacies recognition model.

[0087] Step S106: input the target logging data of the test well point and the target logging phase element value of the test well point into the sedimentary microfacies identification model to obtain the sedimentary microfacies type of the test well point.

[0088] In practical applications, after obtaining the target logging data and target logging phase element values ​​of the well points to be identified, the well points to be identified can be divided into sample well points and test well points. According to the target logging phase element values ​​of the sample well points, the sedimentary microfacies types of the sample well points can be determined. The target logging data of the sample well points and the target logging phase element values ​​of the sample well points are used as input, and the sedimentary microfacies types of the sample well points are used as output to train the deep learning network, and a trained sedimentary microfacies recognition model is obtained, which can identify the sedimentary microfacies types of the well points according to the logging data and logging phase element values ​​of the well points. In this way, the sedimentary microfacies types of the test well points can be obtained by simply inputting the target logging data of the test well points and the target logging phase element values ​​of the test well points into the sedimentary microfacies recognition model.

[0089] In practical applications, the well points to be identified can be randomly divided into sample well points and test well points. For example, the area where all the well points are located can be divided into a certain number of grids, and then a certain proportion of well points in each grid can be randomly selected as sample well points, and the remaining well points can be used as test well points. For example, according to the area of ​​the area where the well points are located, 900m is selected as the step length, and the area is gridded at equal distances, and the area is divided into 29 grids. Assuming that there are an average of 100 wells in each grid in the area, 5% of the well points are randomly selected for each grid, and a total of 143 well points are selected as sample well points, and the remaining well points can be used as test well points. In order to understand the relationship between logging data, logging phase element values ​​and sedimentary microfacies types, the n_1 formation of one of the wells well_1 is taken as an example. The logging phase elements of the four selected logging curves are calculated respectively, and then the logging phase element values ​​of the four logging curves are averaged to obtain the logging phase element value of the n_1 formation of the well point well_1. The amplitude of the logging curve is calculated by the root mean square method, and the amplitude of the logging curve is 0.8, indicating that the amplitude of the curve is large; the shape of the logging curve is calculated by the statistical characteristic method, and its mean Mean: 0.725∈[0.7,1.0], variance V ariance: 0.01∈[0.0,0.1], Skewness: 1.82∈[0.0,10], Kurtosis: 0.005∈[0.0,0.1]. According to the statistical law, the shape of the logging curve is box-type. The smoothness of the logging curve is calculated by the tooth amplitude, and the smoothness value is 0.09, indicating that the curve is relatively smooth. The thickness is calculated by the vertical depth data of the formation, and the thickness is 5.6 meters. Combining the logging phase elements and geological review, it can be determined that the sedimentary microfacies type of the n_1 formation of the well_1 well point is a river channel.

[0090] In a specific application scenario, the target logging data and target logging phase element values ​​of the sample well points are used as input, and the sedimentary microfacies types of the sample well points are used as output. The deep learning network is trained to obtain a trained sedimentary microfacies identification model. The number Ns of sample well points of each type of sedimentary microfacies type s can be determined. The weight of each type of sedimentary microfacies type is determined according to the inverse proportion of the number. The weight generation formula can be: , represents the weight of the sth sedimentary microfacies type, S represents the total number of sedimentary microfacies types, The number of sample well points representing the p-th type of sedimentary microfacies is determined; the initial deep learning network is determined; the target logging data of the sample well points and the target logging phase element values ​​of the sample well points are input into the deep learning network; the prediction probability of the sedimentary microfacies type of the sample well points by the deep learning network is obtained; based on the weight, prediction probability, and sedimentary microfacies type of the sample well points, the loss value of the deep learning network is generated. The loss value generation formula can be , loss represents the loss value, A represents the number of sample well points, q represents the number of sample well points, It represents the true value of the pth category in the category code corresponding to the qth sample. The true value can be represented by 0 or 1. Represents the predicted probability of the deep learning network that the qth sample belongs to the pth category; the deep learning network is updated based on the loss value to obtain a trained sedimentary microfacies recognition model.

[0091] In specific application scenarios, the architecture of the sedimentary microfacies identification model can be flexibly adjusted according to actual needs. For example, it can include a first same convolution block, a second same convolution block, a first downsampling layer, a third same convolution block, a second downsampling layer, a convolution layer, a connection layer, and four fully connected layers connected in sequence; and the same convolution block includes a same convolution layer, a normalization layer (BatchNorm) and an activation layer, and the activation layer can be a Relu activation function, etc.; and the input of the first same convolution block includes the target logging data, and the input of the connection layer includes the output of the convolution layer and the target logging phase element value, such as Figure 3 As shown in the figure, assuming that the data dimension of the input sedimentary microfacies identification model is (B, 1, H, 4), where H is the number of sampling points of the logging data, then after the first convolution block, the second convolution block, the first downsampling layer and the convolution layer, the network output data dimension is (B, 1, 1, 1), and then the output data is reduced to (B, 1, 1). This data is the feature map of the logging curve data. The processed logging phase element data dimension is (B, 4). The connection layer upgrades the logging phase element data to (B, 1, 4) and merges it with the feature map of the logging curve data. The merged dimension is (B, 1, 5). The merged data is input into the subsequent fully connected layer, and finally the sedimentary microfacies type of the well point to be identified is output. The specific data form can be category coding, etc. It should be noted that the size of each layer in the sedimentary microfacies identification model can be flexibly adjusted. For example, the convolution kernels of the first and second convolution blocks can be 7*3, the convolution kernel of the third convolution block can be 5*1, and the convolution kernel of the convolution layer can be 10*1.

[0092] It should be noted that in the process of training the deep learning network, the sedimentary microfacies type of the well point can be encoded, and the encoding method can be One-Hot encoding. The standardized logging data, logging phase element data, and sedimentary microfacies type encoding are organized into a data set. The data set format is [[natural potential], [natural gamma], [micro gradient], [micro potential], [logging curve amplitude, logging curve shape, thickness, curve smoothness], [sedimentary microfacies type]], and then this data set is used to train the deep learning network. And the training process can be determined according to the recognition requirements. For example, when training the model, a batch can be set to 10, the model training is 1000 rounds, and the learning rate is set to 0.001.

[0093] The present application provides a sedimentary microfacies identification method, which comprises the following steps: obtaining target logging data of a well point to be identified; generating target logging facies element values ​​of the well point to be identified based on the target logging data of the well point to be identified; dividing the well point to be identified into sample well points and test well points; determining the sedimentary microfacies type of the sample well point based on the target logging facies element values ​​of the sample well points; training a deep learning network using the target logging data of the sample well point and the target logging facies element values ​​of the sample well point as inputs and the sedimentary microfacies type of the sample well point as outputs to obtain a trained sedimentary microfacies identification model; inputting the target logging data of the test well point and the target logging facies element values ​​of the test well point into the sedimentary microfacies identification model to obtain the sedimentary microfacies type of the test well point. The present application divides the well points to be identified into sample well points and test well points, and obtains a sedimentary microfacies identification model through training of the target logging data, target logging phase element values ​​and sedimentary microfacies types of the sample well points, so that the sedimentary microfacies types of the test well points can be identified with the help of the sedimentary microfacies identification model. No manual operation is required, the degree of automation is high, the identification efficiency of the sedimentary microfacies types is improved, the operational errors caused by manual identification are avoided, and the identification accuracy of the sedimentary microfacies types is improved.

[0094] See also Figure 4 , Figure 4 A schematic diagram of the structure of a sedimentary microfacies identification system provided in an embodiment of the present application.

[0095] A sedimentary microfacies identification system provided in an embodiment of the present application may include:

[0096] The first acquisition module 101 is used to acquire target logging data of a well point to be identified;

[0097] The first generating module 102 is used to generate the target logging phase element value of the well point to be identified according to the target logging data of the well point to be identified;

[0098] The first division module 103 is used to divide the well points to be identified into sample well points and test well points;

[0099] The first determination module 104 is used to determine the sedimentary microfacies type of the sample well point according to the target logging facies element value of the sample well point;

[0100] The first training module 105 is used to take the target logging data of the sample well point and the target logging phase element value of the sample well point as input, and take the sedimentary microfacies type of the sample well point as output, to train the deep learning network and obtain a trained sedimentary microfacies recognition model;

[0101] The first identification module 106 is used to input the target logging data of the test well point and the target logging phase element value of the test well point into the sedimentary microfacies identification model to obtain the sedimentary microfacies type of the test well point.

[0102] In a sedimentary microfacies identification system provided in an embodiment of the present application, a first training module may include:

[0103] A first determination unit is used to determine the number of sample well points of each type of sedimentary microfacies;

[0104] The second determination unit is used to determine the weight of each type of sedimentary microfacies according to the inverse ratio of the quantity;

[0105] A third determining unit, used to determine an initial deep learning network;

[0106] A first input unit is used to input the target logging data of the sample well point and the target logging phase element value of the sample well point into the deep learning network;

[0107] A first acquisition unit is used to obtain the prediction probability of the sedimentary microfacies type of the sample well point by the deep learning network;

[0108] A first generating unit is used to generate a loss value of a deep learning network based on the weight, the prediction probability, and the sedimentary microfacies type of the sample well point;

[0109] The first updating unit is used to update the deep learning network based on the loss value to obtain a trained sedimentary microfacies identification model.

[0110] A sedimentary microfacies identification system provided by an embodiment of the present application, wherein the sedimentary microfacies identification model comprises a first same convolution block, a second same convolution block, a first downsampling layer, a third same convolution block, a second downsampling layer, a convolution layer, a connection layer and four fully connected layers connected in sequence; and the same convolution block comprises a same convolution layer, a normalization layer and an activation layer;

[0111] The input of the first convolutional block includes target logging data, and the input of the connection layer includes the output of the convolutional layer and the target logging phase element value.

[0112] In a sedimentary microfacies identification system provided in an embodiment of the present application, a first acquisition module may include:

[0113] A second acquisition unit is used to acquire target logging data of the well point to be identified, wherein the target logging data includes natural potential, natural gamma, micro gradient and micro electrical rate;

[0114] The first generation module may include:

[0115] The second generating unit is used to generate target logging phase element values ​​of the well point to be identified according to the target logging data of the well point to be identified, and the target logging phase element values ​​include logging curve amplitude, logging curve shape, thickness and curve smoothness.

[0116] In a sedimentary microfacies identification system provided by an embodiment of the present application, the second generation unit can be used to: obtain a set data interval; count the proportion of the target logging data of the well point to be identified in each data interval; sort the target logging data of the well point to be identified according to the size of the proportion to obtain a sorting result; filter out the first type of target logging data with a proportion of the target value from the sorting result; and generate the logging curve amplitude of the well point to be identified based on the first type of target logging data.

[0117] In a sedimentary microfacies identification system provided by an embodiment of the present application, the second generation unit can be used to: convert the target logging data of each well point to be identified into a target feature vector; generate a central value for each type of target logging data; generate a central vector of the target logging data based on the central value; generate the Mahalanobis distance between the target feature vector and the central vector of each well point to be identified; and determine the logging curve morphology of the well point to be identified based on the Mahalanobis distance of the well point to be identified.

[0118] A sedimentary microfacies identification system provided by an embodiment of the present application, wherein the second generation unit can be used to: perform discrete Fourier transform on target logging data of a well point to be identified to obtain initial frequency domain data; determine a frequency threshold based on the initial frequency domain data; in the initial frequency domain data, filter frequency domain data above the frequency threshold as target frequency domain data; perform inverse Fourier transform on the target frequency domain data to obtain second-type target logging data; generate a toothing amplitude for the second-type target logging data; and determine the smoothness of the curve of the well point to be identified based on the toothing amplitude.

[0119] The present application also provides an electronic device and a computer-readable storage medium, both of which have the corresponding effects of the sedimentary microfacies identification method provided in the embodiment of the present application. Figure 5 , Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0120] An electronic device provided in an embodiment of the present application includes a memory 201 and a processor 202. The memory 201 stores a computer program. When the processor 202 executes the computer program, the steps of the sedimentary microphase identification method described in any of the above embodiments are implemented.

[0121] See also Figure 6, another electronic device provided in the embodiment of the present application may also include: an input port 203 connected to the processor 202, for transmitting commands input from the outside to the processor 202; a display unit 204 connected to the processor 202, for displaying the processing results of the processor 202 to the outside; a communication module 205 connected to the processor 202, for realizing communication between the electronic device and the outside. The display unit 204 can be a display panel, a laser scanning display, etc.; the communication method adopted by the communication module 205 includes but is not limited to mobile high-definition link technology (Mobile High-Definition Link, MHL), Universal Serial Bus (Universal Serial Bus, USB), High-Definition Multimedia Interface (High-DefinitionMultimedia Interface, HDMI), wireless connection: wireless fidelity technology (WIreless Fidelity, WiFi), Bluetooth communication technology, low-power Bluetooth communication technology, and communication technology based on IEEE802.11s.

[0122] An embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of the sedimentary microphase identification method described in any of the above embodiments are implemented.

[0123] The computer-readable storage medium involved in this application includes random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs (Compact Disc Read-Only Memory), or any other form of storage medium known in the technical field.

[0124] For the description of the relevant parts of a sedimentary microphase identification system, an electronic device, and a computer-readable storage medium provided in the embodiments of the present application, please refer to the detailed description of the corresponding parts in a sedimentary microphase identification method provided in the embodiments of the present application, which will not be repeated here. In addition, the parts of the above technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive elaboration.

[0125] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0126] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A sedimentary microfacies identification method, characterized in that: include: Acquire target logging data of the well point to be identified; Generate the target logging phase element value of the well point to be identified according to the target logging data of the well point to be identified, including: performing discrete Fourier transform on the target logging data of the well point to be identified to obtain initial frequency domain data; determining a frequency threshold according to the initial frequency domain data; screening frequency domain data higher than the frequency threshold as target frequency domain data in the initial frequency domain data; performing inverse Fourier transform on the target frequency domain data to obtain second-type target logging data; generating a toothing amplitude of the second-type target logging data; and determining the smoothness of the curve of the well point to be identified based on the toothing amplitude; Divide the area where all the well points to be identified are located into a grid, randomly select sample well points from the grid, and use the remaining well points as test well points; Determining the sedimentary microfacies type of the sample well point according to the target logging facies element value of the sample well point; The target logging data of the sample well point and the target logging phase element value of the sample well point are used as input, and the sedimentary microfacies type of the sample well point is used as output, and a deep learning network is trained to obtain a trained sedimentary microfacies identification model, including: performing one-hot encoding on the sedimentary microfacies type of the sample well point, arranging the standardized logging data, logging phase element data, and sedimentary microfacies type encoding into a data set, and using the data set to train the deep learning network; Inputting the target logging data of the test well point and the target logging phase element value of the test well point into the sedimentary microfacies identification model to obtain the sedimentary microfacies type of the test well point; Among them, the target logging data include natural potential, natural gamma, micro gradient and micro electrical rate; the target logging phase element values ​​include logging curve amplitude, logging curve shape, thickness and curve smoothness.

2. The method according to claim 1, characterized in that: The method uses the target logging data of the sample well point and the target logging phase element value of the sample well point as input, and uses the sedimentary microfacies type of the sample well point as output to train the deep learning network to obtain a trained sedimentary microfacies identification model, including: Determining the number of said sample well points for each sedimentary microfacies type; Determine the weight of each sedimentary microfacies type based on the inverse ratio of said quantities; Determine the initial deep learning network; Inputting the target logging data of the sample well point and the target logging phase element value of the sample well point into the deep learning network; Obtaining the predicted probability of the sedimentary microfacies type of the sample well point by the deep learning network; Generate a loss value of the deep learning network based on the weight, the prediction probability, and the sedimentary microfacies type of the sample well point; The deep learning network is updated based on the loss value to obtain a trained sedimentary microfacies identification model.

3. The method according to claim 2, characterized in that The sedimentary microfacies identification model includes a first convolution block, a second convolution block, a first downsampling layer, a third convolution block, a second downsampling layer, a convolution layer, a connection layer and four fully connected layers connected in sequence; and the convolution block includes a convolution layer, a normalization layer and an activation layer; The input of the first convolutional block includes target logging data, and the input of the connection layer includes the output of the convolutional layer and the target logging phase element value.

4. The method according to claim 1, characterized in that: Generating the logging curve amplitude of the well point to be identified according to the target logging data of the well point to be identified, including: Get the set data interval; Counting the proportion of the target logging data of the well point to be identified in each of the data intervals; According to the size of the proportion, the target logging data of the well point to be identified are sorted to obtain a sorting result; Filter out the first type of target logging data whose proportion is the target value from the sorting results; The logging curve amplitude of the well point to be identified is generated according to the first type of target logging data.

5. The method according to claim 1, characterized in that Generating the logging curve shape of the well point to be identified according to the target logging data of the well point to be identified, including: Converting the target logging data of each of the well points to be identified into a target feature vector; Generate the central value of each type of target logging data; generating a center vector of target logging data according to the center value; Generate a Mahalanobis distance between a target feature vector of each well point to be identified and the center vector; The shape of the well logging curve of the well point to be identified is determined according to the Mahalanobis distance of the well point to be identified.

6. A sedimentary microfacies identification system, characterized in that: include: A first acquisition module is used to acquire target logging data of a well point to be identified; The first generation module is used to generate the target logging phase element value of the well point to be identified according to the target logging data of the well point to be identified, including: performing discrete Fourier transform on the target logging data of the well point to be identified to obtain initial frequency domain data; determining a frequency threshold according to the initial frequency domain data; screening frequency domain data higher than the frequency threshold as target frequency domain data in the initial frequency domain data; performing inverse Fourier transform on the target frequency domain data to obtain a second type of target logging data; generating a toothing amplitude of the second type of target logging data; and determining the smoothness of the curve of the well point to be identified based on the toothing amplitude; The first division module is used to divide the area where all the well points to be identified are located into grids, randomly select sample well points from the grids, and use the remaining well points as test well points; A first determination module is used to determine the sedimentary microfacies type of the sample well point according to the target logging facies element value of the sample well point; The first training module is used to take the target logging data of the sample well point and the target logging phase element value of the sample well point as input, and take the sedimentary microfacies type of the sample well point as output, train the deep learning network, and obtain a trained sedimentary microfacies identification model, including: uniquely encoding the sedimentary microfacies type of the sample well point, arranging the standardized logging data, logging phase element data, and sedimentary microfacies type encoding into a data set, and using the data set to train the deep learning network; A first identification module is used to input the target logging data of the test well point and the target logging phase element value of the test well point into the sedimentary microfacies identification model to obtain the sedimentary microfacies type of the test well point; Among them, the target logging data include natural potential, natural gamma, micro gradient and micro electrical rate; the target logging phase element values ​​include logging curve amplitude, logging curve shape, thickness and curve smoothness.

7. The system according to claim 6, characterized in that The first training module includes: A first determining unit is used to determine the number of the sample well points of each type of sedimentary microfacies; A second determining unit is used to determine the weight of each type of sedimentary microfacies according to the inverse ratio of the quantity; A third determining unit, used to determine an initial deep learning network; A first input unit, used to input the target logging data of the sample well point and the target logging phase element value of the sample well point into the deep learning network; A first acquisition unit is used to obtain the prediction probability of the sedimentary microfacies type of the sample well point by the deep learning network; A first generating unit, configured to generate a loss value of the deep learning network based on the weight, the predicted probability, and the sedimentary microfacies type of the sample well point; The first updating unit is used to update the deep learning network based on the loss value to obtain a trained sedimentary microfacies identification model.

8. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the sedimentary microfacies identification method as claimed in any one of claims 1 to 5 when executing the computer program.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the sedimentary microfacies identification method according to any one of claims 1 to 5 are implemented.

Images