Prediction method for effective hydrocarbon source rock development condition and related device

Abstract

The invention discloses an effective hydrocarbon source rock development condition prediction method and a related device, and relates to the geophysical exploration and comprehensive technical field, the method comprises the following steps: capturing spatial feature processing sequence data through a total organic carbon content prediction model, training a strong complex mapping relation between the total organic carbon content and seismic attributes, and determining the effective hydrocarbon source rock development condition; meanwhile, a support vector machine classification model is utilized to analyze a mapping relation between seismic attributes and lithology data of a virtual well in a research area, intelligent prediction of the total organic carbon content and intelligent judgment and identification of lithology are completed respectively, the generalization ability of the model is improved, regional distribution of the total organic carbon content and the lithology can be predicted more effectively, and the prediction accuracy of the total organic carbon content and the lithology is improved. Therefore, high accuracy is kept in a complex geological environment. And finally, determining the thickness and position of the effective hydrocarbon source rock of the target layer section based on the lithology identification result and the total organic carbon content prediction result, thereby realizing accurate evaluation of the development condition of the effective hydrocarbon source rock.

Description

Technical Field

[0001] The present application relates to the field of geophysical exploration and comprehensive technology, and in particular to a method for predicting the development of effective hydrocarbon source rocks and related devices. Background Art

[0002] Whether effective source rocks are developed is an important prerequisite for the accumulation of oil and gas, and is also an important basis for evaluating the oil and gas resource potential in the exploration area. Effective source rocks play an important role in supplying hydrocarbons, and their scale and distribution range also control the size and distribution of oil and gas fields. Therefore, it is necessary to evaluate the source rocks in the exploration area in the early stage of exploration to provide certain theoretical and methodological basis for later oil and gas exploration work.

[0003] Existing technologies use seismic data to make predictions through traditional machine learning models or simple neural network models combined with source rock development characteristics. This can solve the problem of predicting the planar distribution of source rocks in areas with less logging data and analytical data. Although this method can predict the vertical source rock development of a single well to a certain extent, it still faces some challenges in practical applications, such as limited generalization ability and poor noise resistance in complex geological environments, which affects the accuracy of the prediction of source rock development. Summary of the invention

[0004] The purpose of this application is to provide a method and related device for predicting the development of effective hydrocarbon source rocks, which can improve the accuracy of predicting the development of hydrocarbon source rocks.

[0005] To achieve the above objectives, this application provides the following solutions: In a first aspect, the present application provides a method for predicting the development of effective source rocks, comprising: Obtain post-stack seismic data in the study area; Extracting seismic attributes of a virtual well inserted in the study area based on the post-stack seismic data; Inputting the seismic attributes of the virtual well into a trained lithology prediction model, and outputting lithology data of each layer of the virtual well, wherein the lithology prediction model is a support vector machine classification model, and the lithology data includes lithology types and development thicknesses of different lithologies; Inputting the seismic attributes of the virtual well into the trained total organic carbon content prediction model, and outputting the total organic carbon content of each layer of the virtual well, wherein the total organic carbon content prediction model is a CNN-GRU architecture model based on a combination of a convolutional neural network and a gated recurrent unit; Determine the thickness and position of effective source rock according to the lithology data and total organic carbon content of each layer section of the virtual well, wherein the effective source rock is a mudstone layer with a total organic carbon content greater than a threshold value; Based on the thickness and location of the effective source rock, the development of the effective source rock is evaluated.

[0006] In a second aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for predicting the development of effective source rocks as described in the first aspect above.

[0007] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for predicting the development of effective source rocks as described in the first aspect above.

[0008] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method for predicting the development of effective source rocks as described in the first aspect above.

[0009] According to the specific embodiments provided in this application, this application discloses the following technical effects: The present application provides a prediction method and related device for the development of effective source rocks. The method inserts a virtual well in the research area, inputs the seismic attributes of the virtual well into the trained lithology prediction model and the trained total organic carbon content prediction model, obtains the lithology type and total organic carbon content of each layer of the virtual well, and determines the effective source rock based on this. According to the thickness and position of the effective source rock, the development of the effective source rock is evaluated. The accurate identification of lithology in the present application helps to determine the existence and quality of the source rock, and the accurate prediction of TOC (Total Organic Carbon) content is directly related to the hydrocarbon generation capacity of the source rock. Therefore, the accurate prediction of lithology and TOC content in the present application is a key factor in evaluating the quality and hydrocarbon generation potential of the source rock. Therefore, the present application can improve the accuracy of the prediction of the development of the source rock. In addition, the present application adopts the support vector machine classification (SVC) model and the CNN-GRU architecture model to intelligently identify lithology and intelligently predict TOC. With its excellent classification performance, SVC has shown significant advantages in intelligent identification of lithology. By maximizing the edge strategy, the generalization ability of the model is improved, thereby maintaining high accuracy in complex geological environments. On the other hand, the CNN-GRU model combines the advantages of convolutional neural networks (CNN) and gated recurrent units (GRU). It can not only capture spatial features, but also process sequence data. The CNN-GRU model with strong complex mapping capabilities is trained to establish a nonlinear relationship between TOC data and seismic attributes, and more effectively predict the regional distribution of TOC content. Through the multi-attribute prediction of neural networks, this application finally applies the trained mathematical model to three-dimensional seismic data volumes, realizing the three-dimensional scale quantitative evaluation of source rocks. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] 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 will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0011] Figure 1 A schematic flow chart of a method for predicting the development of effective source rocks provided in Example 1 of the present application; Figure 2 This is a flow chart of the training method of the SVC model and the CNN-GRU architecture model in Example 1 of the present application; Figure 3 This is a schematic diagram of the deployment of a virtual well position in a first specific area in Example 1 of the present application; Figure 4This is a schematic diagram of the deployment of the second virtual well positions in the second specific area in Example 1 of the present application; Figure 5 This is a thermal diagram of the lithology correlation in Example 1 of the present application; Figure 6 This is a heat map of TOC correlation in Example 1 of the present application; Figure 7 This is a comparison diagram of the loss function of the lithology prediction model in Example 1 of the present application; Figure 8 A schematic diagram of a confusion matrix for a first specific area in Example 1 of the present application; Fig. 9 This is a schematic diagram of a two-segment confusion matrix of a second specific area in Example 1 of the present application; Fig.10 A schematic diagram of a ROC curve of a section of mudstone in the first specific area in Example 1 of the present application; Fig.11 This is a schematic diagram of the ROC curve of the second section of mudstone in the second specific area in Example 1 of the present application; Fig.12 This is a comparison chart of the loss function of the TOC prediction model in Example 1 of the present application; Fig.13 This is a graph showing the loss function of the TOC prediction model in Example 1 of the present application; Fig.14 A schematic diagram of the predicted planar development thickness of a section of source rock in the first specific area in Example 1 of the present application; Fig.15 This is a schematic diagram of the predicted planar development thickness of the second section of source rock in the second specific area in Example 1 of the present application. DETAILED DESCRIPTION

[0012] Through research, it is found that related technologies have methods for evaluating the development of effective source rocks, for example: through an earthquake prediction method based on the combination of BP and SOM neural networks, this method uses a self-organizing feature mapping neural network to cluster the original earthquake data for preprocessing, and then uses the BP neural network to learn and predict the sample data; or, by using the TOC curve as a constraint, using seismic multi-attribute inversion technology, after multiple stepwise regression and artificial neural network training and other steps, the seismic attributes with a strong correlation with TOC are selected, and the relationship between seismic attributes and TOC is fitted, and then the TOC data body is obtained, and the development of effective source rocks is evaluated on this basis; or, by training a ConvLSTM neural network with strong complex mapping capabilities to establish a nonlinear relationship between TOC content and seismic elastic parameters, so as to effectively predict the regional distribution of TOC content; or, in a data-driven machine learning framework, using geological constraints, drilling logging, logging, geochemistry and prestack seismic data for quantitative seismic characterization of source rocks. The random forest algorithm performs well in small sample data and can effectively predict the distribution of mudstone and total organic carbon content. However, there are challenges in practical applications, such as limited generalization ability of existing methods, poor noise resistance in complex geological environments, and inability to effectively establish nonlinear relationships between sequence data such as TOC data and seismic attributes, which in turn affects the accurate assessment of the development of effective source rocks.

[0013] In this regard, the present application provides a method and related device for predicting the development of effective source rocks to overcome the above-mentioned technical deficiencies.

[0014] 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.

[0015] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0016] Example 1 The prediction method of the effective source rock development provided by the embodiment of the present application can be applied to an application environment including a terminal and a server. The terminal communicates with the server through a network. The data storage system can store data that the server needs to process. The data storage system can be set up separately, integrated on a server, or placed on a cloud or other server. The terminal can send the post-stack seismic data of the research area to be processed to the server. After receiving the post-stack seismic data of the research area to be processed, the server extracts the seismic attributes of the virtual well inserted in the research area according to the post-stack seismic data; inputs the seismic attributes of the virtual well into the trained lithology prediction model, and outputs the lithology data of each layer of the virtual well, wherein the lithology prediction model is a support vector machine classification model; inputs the seismic attributes of the virtual well into the trained total organic carbon content prediction model, and outputs the total organic carbon content of each layer of the virtual well, wherein the total organic carbon content prediction model is a CNN-GRU architecture model based on a combination of a convolutional neural network and a gated recurrent unit; according to the lithology data and total organic carbon content of each layer of the virtual well, the thickness and position of the effective source rock are determined, wherein the effective source rock is a mudstone layer with a total organic carbon content greater than a threshold; according to the thickness and position of the effective source rock, the development of the effective source rock is evaluated. The server can feed back the evaluation result of the development of the effective source rock to the terminal. In addition, in some embodiments, the prediction method of the development of effective source rocks can also be implemented independently by a server or a terminal. For example, the terminal can directly process the post-stack seismic data of the research area to be processed using the prediction method of the development of effective source rocks, or the server can obtain the post-stack seismic data of the research area to be processed from the data storage system and process it using the prediction method of the development of effective source rocks.

[0017] The terminals may be, but are not limited to, various desktop computers, laptops, smart phones, tablet computers, IoT devices and portable wearable devices. IoT devices may be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server may be implemented as an independent server or a server cluster consisting of multiple servers, or a cloud server.

[0018] In an exemplary embodiment, Figure 1 A method for predicting the development of effective source rocks is provided, and the method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to a server as an example for explanation, and includes the following steps 201 to 206. Among them: Step 201, obtaining post-stack seismic data of the study area.

[0019] Step 202: extracting seismic attributes of a virtual well inserted in the study area based on the post-stack seismic data.

[0020] This embodiment is based on the obtained post-stack seismic data. As the number of real wells in the study area is small and cannot meet the prediction accuracy required for plane analysis, virtual wells are inserted in the study block and fixed intervals are set to make them evenly distributed.

[0021] Step 203, input the seismic attributes of the virtual well into the trained lithology prediction model, and output the lithology data of each layer of the virtual well, wherein the lithology prediction model is a support vector machine classification model, and the lithology data includes lithology type and development thickness of different lithologies.

[0022] Step 204, input the seismic attributes of the virtual well into the trained total organic carbon content prediction model, and output the total organic carbon content of each layer of the virtual well, wherein the total organic carbon content prediction model is a CNN-GRU architecture model based on a combination of a convolutional neural network and a gated recurrent unit.

[0023] Step 205, determining the thickness and position of the effective source rock according to the lithology data and the total organic carbon content of each layer section of the virtual well, wherein the effective source rock is a mudstone layer with a total organic carbon content greater than a threshold.

[0024] The TOC threshold is set based on existing research experience, and the mudstone layer sections with TOC greater than the threshold are determined.

[0025] Step 206: Evaluate the development of the effective source rock based on the thickness and location of the effective source rock.

[0026] The effective source rocks were counted and accumulated to obtain the thickness of effective source rocks in each layer of each virtual well. The thickness values ​​and coordinates of the effective source rocks at each virtual well point were imported into the "DF-GVision" software to draw a plane contour map of the effective source rock thickness within the study area, so as to evaluate the development of effective source rocks.

[0027] As an optional implementation, Figure 2 As shown in the figure, the specific training process of the lithology prediction model and the total organic carbon content prediction model includes: (1) Constructing a sample training set, wherein the sample training set includes a lithology sample data set and a total organic carbon content sample data set of real wells in the study area, the lithology sample data set includes lithology-sensitive seismic attributes and sample lithology data, and the total organic carbon content sample data set includes total organic carbon content-sensitive seismic attributes and sample total organic carbon content interpretation data.

[0028] The construction process of the sample training set specifically includes: (1-1) Obtaining sample geological exploration data of the study area, wherein the sample geological exploration data includes sample well location data, sample well trajectory data, sample post-stack seismic data, sample well logging curve data, sample total organic carbon content interpretation data, sample stratigraphic interpretation data and sample core logging data.

[0029] The interpretation software SMI is used to import sample well location data, sample well trajectory data, sample post-stack seismic data, sample well logging curve data, sample TOC interpretation data, sample layer interpretation data and sample core logging data. The sample TOC interpretation data is read to obtain the real well location coordinates with TOC interpretation data in the effective source rock; the lithology data of a single well is determined based on the sample core logging data, and the lithology data includes lithology type and development thickness of different lithologies; the sample post-stack seismic data is preprocessed, including noise and interference removal, and guidance filtering based on structural dip angle.

[0030] The "SMI" software described in this embodiment is also called seismic waveform indication inversion software, which is a high-precision seismic phase-controlled reservoir prediction software independently developed by Beijing Zhongheng Lihua Petroleum Technology Research Institute. The software uses a Chinese interface and has a three-dimensional visualization function. It is suitable for high-precision reservoir prediction of thin interlayers, tight reservoirs and unconventional reservoirs.

[0031] The sample post-stack seismic data described in this embodiment refers to the acquisition of information about underground rock and strata by exciting vibrations on the ground and then recording the waves propagating back to the surface in seismic exploration. The data obtained by such recording is called three-dimensional seismic data. Post-stack is the three-dimensional seismic data obtained after stacking, which helps to enhance effective signals, reduce noise, and improve the signal-to-noise ratio of data.

[0032] (1-2) Determine the sample lithology data of the real well based on the sample core logging data.

[0033] (1-3) Constructing a velocity field based on the sample well location data and the sample logging curve data.

[0034] Based on the sample well location data and sample logging curve data, the velocity data and density data measured by the real well at the same depth point are used to calculate the acoustic impedance of the formation, and then the reflection coefficient sequence R (t) of the formation is calculated. Subsequently, the reflection coefficient sequence is convolved with the seismic wavelet to obtain the synthetic seismic record S (t). The formula is: (1); Where S(t) is the synthetic seismic record; R(t) is the reflection coefficient; W(t) is the seismic wavelet.

[0035] Synthetic seismic records are used to construct time-depth relationships and thus construct velocity fields.

[0036] The seismic wavelet described in this embodiment refers to the waveform of a seismic wave during its underground propagation process. It has limited energy, a definite start time, and a certain continuation length, and is a basic unit in seismic records.

[0037] The synthetic seismic record described in this embodiment refers to seismic record data obtained by combining an actual geological model with the principle of seismic exploration and development.

[0038] The time-depth relationship described in this embodiment refers to converting the time it takes for seismic waves to propagate in underground media into corresponding depth or height relationships in seismic data interpretation.

[0039] The velocity field described in this embodiment is the seismic velocity field, which reflects the spatial variation and distribution of the sound wave propagation velocity of underground rocks and can be used to convert the time domain data on the seismic data into depth domain data.

[0040] (1-4) Obtaining seismic attributes along the well trajectory in the real well depth domain based on the sample well trajectory data, the sample post-stack seismic data, the sample stratigraphic interpretation data and the velocity field.

[0041] Based on sample post-stack seismic data, sample well trajectory data, and sample stratigraphic interpretation data, the time domain seismic volume attributes are extracted; the time domain seismic volume attributes are converted into depth domain seismic volume attributes according to the velocity field; then the required seismic attributes are extracted based on the depth domain seismic volume attributes, and the depth domain seismic attributes along the well trajectory of the real well are extracted based on the stratigraphic interpretation constraints.

[0042] The seismic volume attributes described in this embodiment refer to different three-dimensional attribute volumes calculated according to different algorithms based on the three-dimensional seismic data volume.

[0043] (1-5) A correlation analysis method is used to calculate the correlation coefficients between the seismic attributes along the well trajectory in the real well depth domain and the total organic carbon content and the lithological attributes, respectively, to obtain a first correlation coefficient and a second correlation coefficient.

[0044] Based on the seismic attributes of the real well along the well trajectory, they are preprocessed, including outlier removal, normalization, smoothing, etc. Then, the Pearson correlation coefficient is used to analyze the correlation between TOC and lithology-sensitive real well along the well trajectory, and the correlation is sorted by numerical value. The seismic attributes along the well trajectory in the real well depth domain with the top 80% correlation value between the label data and the seismic attributes along the well trajectory in the real well depth domain with the bottom 80% correlation value between the seismic attributes are selected. The formula is: (2); In the formula, m is the number of sample sizes; r xy is the degree of linear correlation between two variables x and y; is the mean value of variable x; is the mean value of the variable y.

[0045] (1-6) Calculate the correlation coefficient between the seismic attributes in the real well depth domain along the well trajectory to obtain a third correlation coefficient.

[0046] (1-7) Selecting the real well depth domain along-well trajectory seismic attribute for which the first correlation coefficient is greater than a first threshold and the third correlation coefficient is less than a third threshold as the lithology-sensitive seismic attribute.

[0047] (1-8) Selecting the seismic attribute along the well trajectory in the real well depth domain, for which the second correlation coefficient is greater than a second threshold and the third correlation coefficient is less than a third value, as the total organic carbon content sensitive seismic attribute.

[0048] (2) Using the lithology-sensitive seismic attributes as feature data and the sample lithology data as label data to train a lithology prediction model.

[0049] The seismic attributes along the well trajectory in the real well depth domain with the top 80% correlation values ​​with the lithology data and the seismic attributes along the well trajectory in the real well depth domain with the bottom 80% correlation values ​​are selected as feature data, and the lithology data of each single well (i.e., sample lithology data) are encoded using One-hot encoding. The encoded lithology column is used as label data and used together with the feature data as a sample data set. The data set is divided into a training set and a validation set. The support vector machine classification SVC method is used to identify the lithology, and the support vector machine classification SVC method is used to predict the lithology.

[0050] The accuracy_score is used as the evaluation index for the prediction results to evaluate the effect of training. Based on the binary confusion matrix, the ROC curve is made using its data to evaluate the quality of model prediction. The area under the curve is calculated as AUC. The closer the AUC is to 1, the better the model performance; the farther from the 45-degree diagonal line, the higher the accuracy.

[0051] The support vector machine described in this embodiment is a commonly used supervised learning algorithm, which is mainly used for classification and regression analysis. Its basic idea is to separate sample instances of different categories by constructing an optimal hyperplane in the feature space. Lithology classification prediction usually involves multiple geological parameters and features, which may be high-dimensional and complex. The support vector machine model performs well in processing high-dimensional data and complex features, and can effectively perform feature mapping and classification.

[0052] (3) Using the total organic carbon content-sensitive seismic attributes as feature data and the sample total organic carbon content interpretation data as label data to train a total organic carbon content prediction model.

[0053] The seismic attributes along the well trajectory in the real well depth domain with the top 80% correlation values ​​with TOC data and the seismic attributes along the well trajectory in the real well depth domain with the bottom 80% correlation values ​​are selected as feature data, and the TOC interpretation data is used as label data together with the feature data as the TOC prediction sample data set, and the data set is divided into a training set and a validation set. The CNN-GRU architecture model based on the convolutional neural network CNN and the gated recurrent unit GRU is used for TOC regression prediction. In the training neural network model, the loss function based on the mean square error MSE is used as the performance indicator. By calculating the loss of the training set and the loss diagnosis model of the validation set, a complex mapping relationship between input and output is established to achieve TOC prediction of unknown samples of virtual wells.

[0054] The convolutional neural network CNN described in this embodiment is a deep learning model specially designed for processing and analyzing grid-shaped data. The core idea is to extract local features in the input data through the convolution layer and reduce the spatial dimension of the data through the pooling layer, thereby effectively capturing the patterns and structures in the image.

[0055] The gated recurrent unit GRU described in this embodiment is a recurrent neural network proposed to solve problems such as long-term memory and gradient in back propagation. It controls the flow of information through learnable gates, aiming to better capture dependencies with large time step distances in time series.

[0056] The loss function based on mean square error described in this embodiment is an operational function used to measure the degree of difference between the model's predicted value f(dx) and the true value Y. It is a non-negative real-valued function, usually represented by L(Y, f(x)). The smaller the loss function, the better the generalization ability of the model. The loss function is mainly used in the training phase of the model. After each batch of training data is sent to the model, the predicted value is output through forward propagation, and then the loss function calculates the difference between the predicted value and the true value, that is, the loss value. After obtaining the loss value, the model updates each parameter through back propagation to reduce the loss between the true value and the predicted value, so that the predicted value generated by the model moves closer to the true value, thereby achieving the purpose of learning.

[0057] The mean square error represents the difference between the actual value and the predicted value of the reservoir parameter, and the formula is: (3); In the formula, is the number of original sample data; is the actual value; Predict a numeric value for the model.

[0058] This embodiment provides an effective intelligent prediction method for source rocks based on SVC and CNN-GRU. In the modeling process, lithology data and TOC data use nominal data and sequence data as key input parameters for model training and prediction, respectively. The relationship between lithology data and TOC data and seismic attributes is often nonlinear, and the traditional neural network model requires a large amount of data and a complex network structure to effectively model this relationship. This embodiment uses support vector machine classification SVC and CNN-GRU architecture model to perform intelligent identification of lithology and intelligent prediction of TOC. With its excellent classification performance, SVC shows significant advantages in intelligent identification of lithology. It uses unique hot encoding to convert nominal data into numerical data, and improves the generalization ability of the model by maximizing the edge strategy, thereby maintaining high accuracy in complex geological environments. On the other hand, the CNN-GRU model combines the advantages of convolutional neural networks and gated recurrent units, which can not only capture spatial features, but also process sequence data. The CNN-GRU model with strong complex mapping capabilities is trained to establish a nonlinear relationship between TOC data and seismic parameters, which can more effectively predict the regional distribution of TOC content. Through the multi-attribute prediction of the neural network, this embodiment finally applies the trained mathematical model to the three-dimensional seismic data volume, thereby realizing the three-dimensional quantitative evaluation of the source rock.

[0059] This embodiment takes a certain depression study area as an example to specifically describe the specific implementation process of the above-mentioned method for predicting the development of effective source rocks.

[0060] a. Use interpretation software to import well location data, well trajectory data, conventional logging curves, TOC interpretation data, post-stack seismic data, stratum interpretation data and core logging data of a certain depression work area. Read conventional logging curves and core logging data to obtain the real well data information of the first specific area segment and the second specific area segment of the predicted area; count the real well coordinates with lithology and TOC data in the two layers according to the core logging data and TOC interpretation data, and the lithology information includes lithology type and thickness of different lithology development; pre-process the post-stack seismic data, including removing noise and interference, and pre-processing such as guidance filtering based on structural dip angle.

[0061] b. Based on the post-stack seismic data and well location data obtained in step a, virtual wells are inserted in the study block because the number of real wells in the study block is small and cannot meet the prediction accuracy requirements required for plane analysis. According to the seismic layer interpretation results, 1,271 virtual well locations are arranged in the first specific area. The virtual well location deployment results of the first specific area are shown in Figure 2. Figure 3 As shown in Figure 2, 715 virtual wells are arranged in the second section of the second specific area. The virtual well deployment results of the second section of the second specific area are shown in Figure 2. Figure 4 As shown, the virtual wells are evenly distributed and the well spacing is about 2700 meters.

[0062] c. Based on the well location data and conventional logging curves obtained in step a, the velocity data and density data measured at the same depth points of 31 drilled wells (real wells) are used to calculate the acoustic impedance, and then the reflection coefficient R(t) sequence is calculated. Subsequently, the convolution of the formation reflection coefficient sequence and the seismic wavelet is calculated to obtain the synthetic seismic record S(t). The synthetic seismic record is used to construct the time-depth relationship, thereby constructing the velocity field.

[0063] d. Based on the work area data obtained in step a, extract the time domain seismic volume attributes; based on the velocity field established in step c, convert the time domain seismic volume attributes into depth domain seismic volume attributes, and then extract nine attributes including instantaneous amplitude, instantaneous frequency, sweet spot, gradient modulus, variance, root mean square amplitude, instantaneous phase, trace integral, and amplitude square difference based on the depth domain seismic volume attributes, and extract the depth domain seismic attributes along the well trajectory of the real well based on the stratigraphic interpretation constraints.

[0064] e. Based on the real well trajectory seismic attributes obtained in step d, preprocess them, including outlier removal, smoothing, normalization, etc., and then use the Pearson correlation coefficient to perform correlation analysis on TOC and lithology-sensitive seismic attributes, and select characteristic attributes with high correlation with label data and low correlation between attributes.

[0065] The seven attributes most sensitive to lithology, namely, root mean square amplitude, instantaneous amplitude, instantaneous frequency, sweet spot, gradient modulus, variance, and trace integral, are selected and calculated to obtain the correlation heat map between seismic attributes and lithology, such as Figure 5 As shown, it can be seen that the track integral attribute is most correlated with TOC data in this study, and the correlation coefficient can reach 0.1616. Among them, the worst correlation is the instantaneous phase attribute, which is only 0.02765, at the end of the value of the correlation between each attribute and TOC data, and the track integral attribute is highly correlated with the instantaneous phase attribute, up to 0.616. The feature selection process should select feature attributes with high correlation with label data and low correlation between attributes, so the root mean square amplitude, track integral, and amplitude square difference are selected as lithology prediction variables.

[0066] The four attributes most sensitive to TOC, namely, root mean square amplitude, instantaneous phase, trace integral, and amplitude square difference, are selected and calculated to obtain the correlation heat map between seismic attributes and TOC, as shown in Figure 2. Figure 6 As shown, it can be seen that the track integral attribute is the most correlated with the TOC data in this study, and the correlation coefficient can reach 0.1616. Among them, the worst correlation is the instantaneous phase attribute, which is only 0.02765, at the end of the numerical value of the correlation between each attribute and the TOC data, and the track integral attribute is highly correlated with the instantaneous phase attribute, up to 0.616. The feature selection process should select feature attributes with high correlation with the label data and low correlation between each attribute, so the root mean square amplitude, track integral, and amplitude square difference are selected as TOC prediction variables.

[0067] f. The well trajectory attributes that are sensitive to lithology after optimization and preprocessing obtained in step e are used as feature data, and the lithology data read in step a are encoded using the One-hot encoding method (see Table 1). The encoded lithology column is used as label data and used together with the feature data as a sample data set.

[0068] Table 1 Lithology coding data table

[0069] The performance evaluation of the model usually depends on the degree of optimization of the loss function. The decreasing trend of the loss function is an intuitive indicator of the convergence of the model training. Ideally, the loss value should decrease monotonically with the increase of training rounds, reflecting the gradual adaptation of the model to the data distribution and the gradual reduction of errors. If the loss function decreases abnormally or oscillates, it may indicate that the learning rate is not properly selected or the model is overfitting or underfitting.

[0070] The support vector machine model performs particularly well in the lithology prediction task. By comparing and analyzing it with traditional neural network models such as feedforward neural network (FNN), convolutional neural network, long short-term memory network (LSTM) and self-organizing map (SOM), we can gain insight into the performance differences from the changes in the model's loss function. The cross entropy loss function (CELF) is selected as the loss function for this multi-classification data calculation.

[0071] In the initial 50 iterations of training, the loss of the SVC model on the test set is significantly lower than that of other models, such as Figure 7 This reveals that SVC can capture data features more accurately when dealing with lithology prediction problems, thus having a clear advantage in prediction accuracy. This feature of the SVC model is mainly due to the fact that it optimizes the classification boundary by selecting support vectors while maximizing the boundary interval.

[0072] In contrast, although the FNN model has a simple structure and is easy to implement, its performance in processing complex lithology data may be limited by its lack of in-depth understanding of data features. The CNN model has achieved great success in the field of image processing, but its application in sequence data processing may not be as good as models designed specifically for such data. The LSTM model processes time series data by introducing a gating mechanism. Although it can effectively capture long-term dependencies, it may face the risk of overfitting in some cases due to the high complexity of the model. As an unsupervised learning method, the SOM model has unique advantages in data visualization and cluster analysis, but in tasks such as lithology prediction that require accurate classification, its performance may not be as good as the supervised learning SVC model.

[0073] Therefore, the optimized support vector machine classification model is selected to predict lithology. The prediction results are evaluated using accuracy_score as the evaluation index to evaluate the effect of training, and based on the binary confusion matrix (such as Figure 8 and Fig. 9 As shown in the figure), use the data to make a ROC curve diagram (as shown in the figure Fig.10 and Fig.11 As shown in the figure, the horizontal axis is FPR, which indicates the proportion of samples predicted as positive but actually negative to all negative samples; the vertical axis is TPR, which indicates the proportion of samples predicted as positive and actually positive to all positive samples. The area under the ROC curve is calculated as AUC. The larger the AUC, the better the prediction effect. Combining the ROC curve and AUC data, the model prediction effect can be improved by adjusting model parameters and adding features.

[0074] Depend on Figure 8-Figure 11 It can be seen that the average TPR of each lithology in the first specific area group is 97.3%, and the AUC reaches 0.97; the average TPR of the second section of the second specific area group is 96.9%, and the AUC reaches 0.99. It can be seen that the support vector machine model has a high accuracy rate in this study and a good training effect. It can make a relatively accurate prediction of the overall lithology classification based on the existing lithology sample data.

[0075] g. Use the preprocessed TOC-sensitive well trajectory attributes obtained in step e as feature data, use the TOC interpretation data as label data, and use them together with the feature data as the TOC prediction sample data set, and divide the data set into a training set and a validation set in a ratio of 8:2.

[0076] The CNN-GRU model performs particularly well in the lithology prediction task. By comparing and analyzing with traditional neural network models such as multi-layer perceptron (MLP), radial basis neural network (RBF), convolutional neural network, and the combination of long short-term memory network and convolutional neural network (CNN-LSTM), we can gain insight into the performance differences from the changes in the loss function of the model. Since the predicted data is sequence data, the root mean square error (RMSE) is selected as the loss function for calculation. The comparison of the loss functions of the five models is shown in the figure below. Fig.12 As shown in the figure, the CNN-GRU model shows significant prediction accuracy and training efficiency compared to traditional architectures such as MLP, RBF, CNN, and CNN-LSTM.

[0077] During the first 50 rounds of training, the loss curve of the CNN-GRU model showed the most robust downward trend, which indicates its superior performance in time series prediction tasks. The CNN-GRU model combines the efficient extraction of spatial features by convolutional neural networks with the ability of gated recurrent units to capture temporal dependencies, enabling it to achieve more accurate predictions when processing complex spatiotemporal data. The MLP model is often limited in dealing with nonlinear problems due to its simple structure, while the RBF network faces challenges of overfitting and computational efficiency due to its large parameter size. CNN effectively extracts spatial features through convolution and pooling layers and is suitable for image processing tasks. Although the CNN-LSTM model has been enhanced in processing spatiotemporal data, its number of parameters and computational complexity have also increased. In contrast, the CNN-GRU model uses GRU units to reduce model parameters while maintaining the ability to capture long-term dependencies, which makes CNN-GRU more efficient and accurate in sequence data analysis.

[0078] Therefore, the CNN-GRU architecture model based on the combination of convolutional neural network CNN and gated recurrent unit GRU is selected for TOC regression prediction, and prediction work is carried out on the sample data set.

[0079] In the training of the neural network model, the loss function based on the mean square error is used as the performance indicator to calculate the loss of the training set and the loss diagnosis model of the validation set. Through the hyperparameter search function, the optimal parameters of the neural network model are found, and a complex mapping relationship between input and output is established. After 20 rounds of learning, both loss curves begin to converge, and the mean square error gradually decreases until the mean square error reaches the minimum value, and the loss curve achieves a nearly perfect fit (such as Fig.13 As shown in the figure, the MSE value of the model training set is 0.0285, and the MSE value of the model test set is 0.0275, indicating that the model can well predict the TOC distribution in the study area and can realize the TOC prediction of unknown samples in virtual wells.

[0080] Based on the lithology prediction results obtained in step f and the TOC prediction results obtained in step g, the TOC threshold of the first section of the first specific area group is set to 1%, and the TOC threshold of the second section of the second specific area group is set to 2% based on existing research experience. The thickness of mudstone with TOC greater than the threshold in the two layers is counted and accumulated to obtain the effective source rock thickness of each layer of each virtual well. The effective source rock thickness value and coordinates at each virtual well point are imported into the "DF-GVision" software to draw a plane contour map of the effective source rock thickness within the study area, and the real wells in the study area are projected onto the prediction plane map for observation, such as Figure 14-15 The planar development map of the effective source rock thickness predicted by the virtual well shown in the figure is basically consistent with the actual statistical results of the well and the results of the virtual well prediction plan map, which shows the generalization ability and prediction accuracy of this prediction and provides theoretical and practical guidance for the next step of oil and gas exploration.

[0081] The present application also provides an application scenario, which applies the above-mentioned method for predicting the development of effective source rocks. Specifically: The method for predicting the development of effective source rocks provided in this embodiment can be applied in the exploration scenario of oil and gas field development. The scenario includes a geological model establishment link, a data analysis link, and a resource assessment link. The method for predicting the development of effective source rocks provided in this embodiment is a key technology in the data analysis link. This method can help the exploration team identify potential high-yield areas, improve the success rate and economic benefits of resource development, and thus optimize the development strategy of oil and gas fields.

[0082] Example 2 This embodiment provides a computer device, which can be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data in the prediction method of the development of effective source rocks. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a prediction method for the development of an effective source rock in Example 1 is implemented.

[0083] Example 3 This embodiment provides a computer device, including a memory and a processor. The memory stores a computer program. When the processor executes the computer program, a method for predicting the development of effective source rocks in the above-mentioned embodiment 1 is implemented.

[0084] Example 4 This embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a method for predicting the development of effective source rocks in the above-mentioned embodiment 1.

[0085] Example 5 This embodiment provides a computer program product, including a computer program. When the computer program is executed by a processor, the method for predicting the development of effective source rocks in the above-mentioned embodiment 1 is implemented.

[0086] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0087] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0088] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0089] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0090] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for predicting the development of effective source rocks, characterized in that: The method for predicting the development of effective source rocks includes: Obtain post-stack seismic data in the study area; Extracting seismic attributes of a virtual well inserted in the study area based on the post-stack seismic data; Inputting the seismic attributes of the virtual well into a trained lithology prediction model, and outputting lithology data of each layer of the virtual well, wherein the lithology prediction model is a support vector machine classification model, and the lithology data includes lithology types and development thicknesses of different lithologies; Inputting the seismic attributes of the virtual well into the trained total organic carbon content prediction model, and outputting the total organic carbon content of each layer of the virtual well, wherein the total organic carbon content prediction model is a CNN-GRU architecture model based on a combination of a convolutional neural network and a gated recurrent unit; Determine the thickness and position of effective source rock according to the lithology data and total organic carbon content of each layer section of the virtual well, wherein the effective source rock is a mudstone layer with a total organic carbon content greater than a threshold value; The development of the effective source rock is evaluated based on the thickness and location of the effective source rock.

2. The method for predicting the development of effective source rocks according to claim 1, characterized in that: The training process of the lithology prediction model and the total organic carbon content prediction model specifically includes: Constructing a sample training set, wherein the sample training set includes a lithology sample data set and a total organic carbon content sample data set of real wells in the research area, the lithology sample data set includes lithology-sensitive seismic attributes and sample lithology data, and the total organic carbon content sample data set includes total organic carbon content-sensitive seismic attributes and sample total organic carbon content interpretation data; Using the lithology-sensitive seismic attributes as feature data and the sample lithology data as label data to train a lithology prediction model; The total organic carbon content sensitive seismic attributes are used as feature data, and the sample total organic carbon content interpretation data are used as label data to train a total organic carbon content prediction model.

3. The method for predicting the development of effective source rocks according to claim 2, characterized in that: The construction process of the sample training set specifically includes: Acquire sample geological exploration data of the study area, wherein the sample geological exploration data includes sample well location data, sample well trajectory data, sample post-stack seismic data, sample well logging curve data, sample total organic carbon content interpretation data, sample stratigraphic interpretation data and sample core logging data; Determine the sample lithology data of the real well according to the sample core logging data; Constructing a velocity field according to the sample well location data and the sample logging curve data; According to the sample well trajectory data, the sample post-stack seismic data, the sample horizon interpretation data and the velocity field, seismic attributes along the well trajectory in the real well depth domain are obtained; The correlation coefficients of the seismic attributes along the well trajectory in the real well depth domain with the total organic carbon content and the lithological attributes are calculated by using a correlation analysis method, and a first correlation coefficient and a second correlation coefficient are obtained correspondingly; Calculating the correlation coefficient between the seismic attributes along the well trajectory in the real well depth domain to obtain a third correlation coefficient; Selecting the real well depth domain along-well trajectory seismic attribute for which the first correlation coefficient is greater than a first threshold value and the third correlation coefficient is less than a third threshold value as the lithology-sensitive seismic attribute; The seismic attribute along the well trajectory in the real well depth domain, for which the second correlation coefficient is greater than a second threshold value and the third correlation coefficient is less than a third threshold value, is selected as the total organic carbon content sensitive seismic attribute.

4. The method for predicting the development of effective source rocks according to claim 3, characterized in that: Constructing a velocity field according to the sample well location data and the sample logging curve data specifically includes: According to the sample well location data and the sample logging curve data, velocity data and density data of a real well at the same depth point are obtained; Calculating the acoustic impedance of the formation based on the velocity data and density data; Calculating the reflection coefficient of the formation according to the acoustic impedance; Convolution operation is performed on the reflection coefficient and the seismic wavelet to generate a synthetic seismic record; The velocity field is constructed from the synthetic seismic record.

5. The method for predicting the development of effective source rocks according to claim 3, characterized in that: According to the sample well trajectory data, the sample post-stack seismic data, the sample horizon interpretation data and the velocity field, seismic attributes along the well trajectory in the real well depth domain are obtained, specifically including: Extracting real well time domain seismic attributes according to the sample well trajectory data, the sample post-stack seismic data and the sample horizon interpretation data; According to the velocity field, converting the real well time domain seismic attributes into real well depth domain seismic attributes; According to the seismic attributes in the real well depth domain and the sample horizon interpretation data, the seismic attributes along the well trajectory in the real well depth domain are obtained.

6. The method for predicting the development of effective source rocks according to claim 1, characterized in that: According to the thickness and location of the effective source rock, the development of the effective source rock is evaluated, including: Accumulating the thickness of the effective source rock to obtain the thickness of the effective source rock in each layer of each virtual well; According to the thickness and position of each effective source rock, a numerical plane contour map of the thickness of the effective source rock within the research area is drawn; The development of the effective source rock is evaluated based on the numerical plane contour map of the effective source rock thickness.

7. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for predicting the development of effective source rocks as described in any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the development of effective source rocks described in any one of claims 1 to 6 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for predicting the development of effective source rocks described in any one of claims 1 to 6 is implemented.

Images