Logging sedimentary microfacies identification method and device based on CKN-LSTM, equipment and medium

Abstract

The invention discloses a logging sedimentary microfacies identification method and device based on CKN-LSTM, equipment and a medium, and relates to the technical field of computers, and the method comprises the steps: processing logging curve data, and determining a training set and a verification set based on the processed logging curve data and sedimentary microfacies interpretation data; an initial logging sedimentary microfacies recognition model with logging curve data as input is constructed based on CKN-LSTM, model training and iterative optimization are conducted on the initial logging sedimentary microfacies recognition model through the training set and the verification set till model training operation is completed, and a target logging sedimentary microfacies recognition model is determined; and performing data processing, feature extraction and time sequence modeling on to-be-processed logging curve data in sequence based on the target logging sedimentary microfacies identification model so as to determine a sedimentary microfacies identification result corresponding to the to-be-processed logging curve data. According to the invention, the accuracy and precision of identification are improved, and the adaptability to a complex sedimentary environment is enhanced.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and particularly to a method, device, equipment and medium for identifying logging sedimentary microfacies based on CKN-LSTM. Background Art

[0002] Due to the complex underground geological conditions, logging data has characteristics such as non-linearity, time-variation and noise interference. Traditional methods have great limitations in terms of accuracy, adaptability and computational efficiency.

[0003] To solve the above problems, machine learning and deep learning have been introduced into the existing sedimentary microfacies identification solutions. However, there are still some technical bottlenecks in these identification solutions, which limit the actual application effects. Moreover, since the input of the sedimentary microfacies identification model in the existing sedimentary microfacies identification solutions mainly relies on the time series information of logging curves and their activation functions are fixed, this may lead to spatial inconsistency in the identification results of the model in complex sedimentary environments, affecting the overall division effect of microfacies. At the same time, most existing solutions often use the image processing of logging curves and then input them into the CNN network (Convolutional Neural Network) to extract features. However, this process may lead to the loss of details of logging data, affecting the accuracy of the model in identifying sedimentary microfacies. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a method, device, equipment and medium for identifying logging sedimentary microfacies based on CKN-LSTM, which can effectively improve the anti-noise ability, feature extraction ability and generalization performance of the logging sedimentary microfacies identification model, improve the accuracy and precision of identification, enhance the adaptability to complex sedimentary environments, and reduce the computational complexity. The specific solutions are as follows:

[0005] In a first aspect, the present application provides a method for identifying logging sedimentary microfacies based on CKN-LSTM, including:

[0006] Collect logging curve data and sedimentary microfacies interpretation data corresponding to the logging curve data;

[0007] Process the logging curve data based on a preset data processing rule, and construct samples based on the determined processed logging curve data and the sedimentary microfacies interpretation data to determine a target sedimentary microfacies identification training set and a target sedimentary microfacies identification verification set;

[0008] Build an initial logging sedimentary microfacies identification model with logging curve data as the input based on CKN-LSTM, and use the target sedimentary microfacies identification training set and the target sedimentary microfacies identification verification set to train and iteratively optimize the initial logging sedimentary microfacies identification model until the model training operation is completed, and then determine the target logging sedimentary microfacies identification model; the CKN-LSTM is a neural network constructed by introducing the KAN network architecture into the CNN network to replace the original activation function and combining with the LSTM network;

[0009] Based on the target logging sedimentary microfacies identification model, perform data processing, feature extraction, and time series modeling on the logging curve data to be processed in sequence to determine the sedimentary microfacies identification result corresponding to the logging curve data to be processed.

[0010] Optionally, the collecting the logging curve data and the sedimentary microfacies interpretation data corresponding to the logging curve data includes:

[0011] Collect the logging curve data based on the first preset data source;

[0012] Collect the sedimentary microfacies interpretation data corresponding to the logging curve data based on the second preset data source; the sedimentary microfacies interpretation data includes core analysis data.

[0013] Optionally, the processing the logging curve data based on the preset data processing rules includes:

[0014] Perform outlier detection on the logging curve data based on the preset outlier detection strategy to determine the first outlier detection result;

[0015] When the first outlier detection result indicates that there are outliers in the logging curve data, trigger the corresponding first outlier processing operation based on the preset outlier processing strategy;

[0016] After completing the first outlier processing operation, determine the target correlation coefficient value between different logging curves by analyzing the logging curve data, and perform logging curve screening based on the target correlation coefficient value to obtain the first screening result;

[0017] Perform standardization processing on the logging curve data in the first screening result based on the preset standard layer feature information to determine the processed logging curve data.

[0018] Optionally, the constructing samples based on the determined processed logging curve data and the sedimentary microfacies interpretation data to determine the target sedimentary microfacies identification training set and the target sedimentary microfacies identification verification set includes:

[0019] Generate sedimentary microfacies labels corresponding to the processed well logging curve data based on the sedimentary microfacies interpretation data and a preset expert annotation interface;

[0020] Annotate the processed well logging curve data based on the sedimentary microfacies labels and preset annotation rules to determine target samples;

[0021] Divide the target samples based on a preset data ratio to determine a target sedimentary microfacies identification training set, a target sedimentary microfacies identification validation set, and a target sedimentary microfacies identification test set.

[0022] Optionally, the data processing, feature extraction, and time series modeling of the well logging curve data to be processed by the target well logging sedimentary microfacies identification model include:

[0023] After completing the model evaluation operation of the target well logging sedimentary microfacies identification model based on the target sedimentary microfacies identification test set and the model evaluation result meets the preset conditions, receive the well logging curve data to be processed through the target well logging sedimentary microfacies identification model;

[0024] Perform outlier detection on the well logging curve data to be processed through the CKN network in the target well logging sedimentary microfacies identification model and the preset outlier detection strategy to determine the second outlier detection result; the CKN network is a network obtained by introducing the KAN network architecture into the CNN network to replace the original activation function;

[0025] When the second outlier detection result indicates that there are outliers in the well logging curve data to be processed, trigger the corresponding second outlier processing operation through the CKN network and based on the preset outlier processing strategy;

[0026] After completing the second outlier processing operation, perform correlation analysis on the well logging curve data to be processed through the CKN network and a plurality of preconfigured correlation coefficients respectively to determine the correlation analysis result;

[0027] Display the correlation analysis result through the CKN network and a preset result drawing and display strategy to complete the well logging curve screening operation and determine the second screening result;

[0028] Perform normalization processing on the second screening result through the CKN network and based on preset standard layer feature information to complete the data processing operation corresponding to the well logging curve data to be processed and determine the processed target well logging curve data;

[0029] Extract features from the target well logging curve data through the CKN network to determine the feature extraction result;

[0030] Perform time series modeling through the LSTM network in the target logging sedimentary microfacies identification model and the feature extraction result to determine the sedimentary microfacies identification result corresponding to the logging curve data to be processed.

[0031] Optionally, the standardizing the second screening result through the CKN network and based on the preset standard layer feature information includes:

[0032] Select a standard layer for the second screening result through the CKN network and based on the preset standard layer feature information to determine the target standard layer;

[0033] Draw a frequency distribution histogram of the curves of all wells in the target standard layer through the target standard layer and the second screening result, and perform data analysis and correction based on the frequency distribution histogram to complete the standardizing operation.

[0034] Optionally, the triggering of the corresponding second outlier processing operation through the CKN network and based on the preset outlier processing strategy includes:

[0035] Delete the outliers in the logging curve data to be processed based on the second outlier detection result;

[0036] Or, determine the mean of the first target data segment corresponding to the outliers in the logging curve data to be processed through the mean generation rule, and perform outlier replacement based on the mean;

[0037] Or, determine the median of the second target data segment corresponding to the outliers in the logging curve data to be processed through the median determination rule, and perform outlier replacement based on the median.

[0038] In a second aspect, the present application provides a logging sedimentary microfacies identification device based on CKN-LSTM, including:

[0039] A data collection module, configured to collect logging curve data and sedimentary microfacies interpretation data corresponding to the logging curve data;

[0040] A sample construction module, configured to process the logging curve data based on a preset data processing rule, and perform sample construction based on the determined processed logging curve data and the sedimentary microfacies interpretation data to determine a target sedimentary microfacies identification training set and a target sedimentary microfacies identification verification set;

[0041] A model training module, which is used to build an initial logging sedimentary microfacies identification model with logging curve data as input based on CKN-LSTM, and use the target sedimentary microfacies identification training set and the target sedimentary microfacies identification validation set to perform model training and iterative optimization on the initial logging sedimentary microfacies identification model until the model training operation is completed, and determine the target logging sedimentary microfacies identification model; the CKN-LSTM is a neural network constructed by introducing the KAN network architecture in the CNN network to replace the original activation function and combining with the LSTM network;

[0042] A sedimentary microfacies identification module, which is used to perform data processing, feature extraction and time series modeling on the to-be-processed logging curve data in sequence based on the target logging sedimentary microfacies identification model to determine the sedimentary microfacies identification result corresponding to the to-be-processed logging curve data.

[0043] In a third aspect, the present application provides an electronic device, including:

[0044] A memory, which is used to store a computer program;

[0045] A processor, which is used to execute the computer program to implement the steps of the foregoing logging sedimentary microfacies identification method based on CKN-LSTM.

[0046] In a fourth aspect, the present application provides a computer-readable storage medium, which is used to store a computer program, and when the computer program is executed by a processor, the steps of the foregoing logging sedimentary microfacies identification method based on CKN-LSTM are implemented.

[0047] It can be seen that in this application, well logging curve data and sedimentary microfacies interpretation data corresponding to the well logging curve data are collected; the well logging curve data is processed based on a preset data processing rule, and a sample is constructed based on the determined processed well logging curve data and the sedimentary microfacies interpretation data to determine a target sedimentary microfacies recognition training set and a target sedimentary microfacies recognition verification set; an initial well logging sedimentary microfacies recognition model with the well logging curve data as the input is constructed based on CKN-LSTM, and the initial well logging sedimentary microfacies recognition model is trained and iteratively optimized using the target sedimentary microfacies recognition training set and the target sedimentary microfacies recognition verification set until the model training operation is completed, and a target well logging sedimentary microfacies recognition model is determined; the CKN-LSTM is a neural network constructed by introducing a KAN network architecture into the CNN network to replace the original activation function and combining it with the LSTM network; based on the target well logging sedimentary microfacies recognition model, the well logging curve data to be processed is sequentially subjected to data processing, feature extraction, and time series modeling to determine the sedimentary microfacies recognition result corresponding to the well logging curve data to be processed. That is to say, in the CNN network of this application, the original activation function is replaced with a KAN network, and an initial well logging sedimentary microfacies recognition model with the well logging curve data as the input is constructed by combining it with the LSTM network. And the collected well logging curve data and sedimentary microfacies interpretation data are processed to determine a target sedimentary microfacies recognition training set. Then, the initial well logging sedimentary microfacies recognition model is trained using the target sedimentary microfacies recognition training set. After the training is completed, based on the determined target well logging sedimentary microfacies recognition model, the well logging curve data to be processed is directly subjected to data processing, feature extraction, and time series modeling in sequence to determine the sedimentary microfacies recognition result. In this way, the anti-noise ability, feature extraction ability, and generalization performance of the well logging sedimentary microfacies recognition model can be effectively improved, the accuracy and precision of recognition can be improved, the adaptability to complex sedimentary environments can be enhanced, and the computational complexity can be reduced. Description of the Drawings

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on the provided drawings without creative efforts.

[0049] Figure 1 Flowchart of a well logging sedimentary microfacies recognition method based on CKN-LSTM provided by this application;

[0050] Figure 2 Schematic diagram of the outlier detection result by the box plot method provided by this application;

[0051] Figure 3 A schematic diagram of the correlation analysis results of logging curves provided by this application;

[0052] Figure 4 A schematic diagram of the network architecture of CKN-LSTM provided by this application;

[0053] Figure 5 A schematic diagram of the network architecture of KAN provided by this application;

[0054] Figure 6 A schematic diagram of the network architecture of LSTM provided by this application;

[0055] Figure 7 A comparison schematic diagram of the identification results of logging sedimentary microfacies provided by this application;

[0056] Figure 8 A schematic diagram of the structure of a logging sedimentary microfacies identification device based on CKN-LSTM provided by this application;

[0057] Figure 9 A structural diagram of an electronic device provided by this application. Detailed implementation manners

[0058] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0059] In the existing sedimentary microfacies identification solutions, machine learning and deep learning have been introduced, but there are still some technical bottlenecks in their identification solutions, which limit the actual application effects. However, since the input of the sedimentary microfacies identification model in the existing sedimentary microfacies identification solutions mainly depends on the time series information of logging curves, and its activation functions are fixed, this may lead to spatial inconsistency in the identification results of the model in complex sedimentary environments, affecting the overall division effect of microfacies. At the same time, most existing solutions often use the graphical processing of logging curves and then input them into a CNN network (Convolutional Neural Network) to extract features, but this process may cause the loss of logging data details and affect the accuracy of the model in identifying sedimentary microfacies. Therefore, this application provides a logging sedimentary microfacies identification solution based on CKN-LSTM, which can effectively improve the noise resistance, feature extraction ability, and generalization performance of the logging sedimentary microfacies identification model.

[0060] See Figure 1As shown in the figure, an embodiment of the present invention discloses a logging sedimentary microfacies identification method based on CKN-LSTM, including:

[0061] Step S11: Collect logging curve data and sedimentary microfacies interpretation data corresponding to the logging curve data.

[0062] Specifically, in this embodiment, relevant data needs to be collected first, that is, collect logging curve data based on the first preset data source; collect sedimentary microfacies interpretation data corresponding to the logging curve data based on the second preset data source; the sedimentary microfacies interpretation data includes core analysis data. Among them, the first preset data source and the second data source can be pre-configured based on actual needs or can be customized and adjusted. It should be understood that the collected logging curve data is a logging curve that can provide physical information of rock formations, such as natural gamma, acoustic time difference, resistivity, etc., and the sedimentary microfacies interpretation data is data combined with core analysis.

[0063] Step S12: Process the logging curve data based on preset data processing rules, and construct samples based on the determined processed logging curve data and the sedimentary microfacies interpretation data to determine the target sedimentary microfacies identification training set and the target sedimentary microfacies identification verification set.

[0064] Specifically, in this embodiment, to ensure the accuracy of the data and lay a solid foundation for model training, the collected data needs to be processed. That is, first, perform outlier detection on the logging curve data based on a preset outlier detection strategy to determine the first outlier detection result; then when the first outlier detection result indicates that there are outliers in the logging curve data, trigger a corresponding first outlier processing operation based on a preset outlier processing strategy; after completing the first outlier processing operation, determine the target correlation coefficient value between different logging curves by analyzing the logging curve data, and perform logging curve screening based on the target correlation coefficient value to obtain the first screening result; then perform standardization processing on the logging curve data in the first screening result based on preset standard layer feature information to determine the processed logging curve data. It can be understood that the outliers in the logging curve data can be manifested as mutation points, extreme values or unreasonable fluctuations in the curve. Combined with Figure 2 As shown in the figure (the vertical coordinate in the figure represents the natural gamma logging curve (which is a logging curve, and the unit of the curve value is API)), the preset outlier detection strategy in this embodiment can be an outlier detection strategy based on the box plot method, and the normal range can be determined by calculating the interquartile range, and the data points outside this range are regarded as outliers. And after detecting outliers, the processing methods include deleting outlier data points and replacing them with reasonable estimated values, such as linear interpolation.

[0065] In addition, in the network part for outlier processing, custom detection thresholds and processing strategies are also supported, and the outlier processing scheme is optimized in combination with the geological background to ensure the accuracy and reliability of subsequent analysis.

[0066] It can be further understood that after the outlier processing is completed, in combination with Figure 3 As shown, in this embodiment, it is also necessary to use three correlation coefficients, Pearson, Spearman, and Kendall, to perform correlation analysis on different logging curves, comprehensively measure the linear and non-linear relationships between variables, determine the magnitudes of the correlation coefficients between different logging curves, and generate a heat map for visualization. In this way, the correlation matrix obtained by calculation intuitively shows the degree of association between logging curves through the heat map, so as to screen out the logging curves that are most sensitive to sedimentary microfacies identification. Among them, Figure 3 RLLS in it represents the shallow lateral resistivity logging curve, GR represents the natural gamma logging curve, SP represents the spontaneous potential logging curve, AC represents the acoustic logging curve, DEN represents the density logging curve, and CNL represents the compensated neutron logging curve.

[0067] After the screening is completed, logging curve standardization processing is carried out. First, select a standard layer section of sedimentation that is sedimentologically stable, has a certain thickness, obvious lithology and logging response characteristics, and is widely distributed as the standard layer. Then, based on the standard layer, draw the frequency distribution histogram of the curves of each single well in the whole area at the standard layer. Then, use the peak value of the curve distribution of the key wells in the figure as the standard to correct other wells to complete the standardization processing and determine the processed logging curve data.

[0068] Furthermore, after the data processing is completed, in order to enable the processed logging curve data to effectively act on model training, some related operations also need to be performed on it. That is, based on the sedimentary microfacies interpretation data and the preset expert annotation interface, generate sedimentary microfacies labels corresponding to the processed logging curve data; based on the sedimentary microfacies labels and the preset annotation rules, annotate the processed logging curve data to determine the target samples; based on the preset data ratio, divide the target samples to determine the target sedimentary microfacies identification training set, the target sedimentary microfacies identification verification set, and the target sedimentary microfacies identification test set. That is to say, in this embodiment, based on the standardized logging curve data, a labeled sedimentary micro-intelligent facies identification sample set is constructed and divided into a training set, a verification set, and a test set to support the construction and evaluation of the deep learning model. First, in combination with the geological expert annotation results, sedimentary microfacies labels are assigned to the standardized data to ensure the geological significance and classification accuracy of the data. Then, according to the ratio, for example, 70% for the training set, 15% for the verification set, and 15% for the test set, the sample set is divided.

[0069] Step S13: Build an initial logging sedimentary microfacies identification model with logging curve data as the input based on CKN-LSTM, and use the target sedimentary microfacies identification training set and the target sedimentary microfacies identification validation set to train and iteratively optimize the initial logging sedimentary microfacies identification model. When the model training operation is completed, determine the target logging sedimentary microfacies identification model; CKN-LSTM is a neural network constructed by introducing the KAN network architecture into the CNN network to replace the original activation function and combining with the LSTM network.

[0070] It should be understood that for the model constructed in this embodiment, the model input is the logging curve data without data processing. As shown in Figure 4 The model architecture is the CKN-LSTM proposed in this embodiment, which is a neural network constructed by introducing the KAN network architecture (Kolmogorov-Arnold Network, a new type of neural network architecture, as shown in Figure 5 ) into the CNN network (Convolutional Neural Network) to replace the original activation function and combining with the LSTM network (Long Short-Term Memory, as shown in Figure 6 ). After completing the initial model construction, the model can be trained based on the determined target sedimentary microfacies identification training set, target sedimentary microfacies identification validation set, and target sedimentary microfacies identification test set. The training set is used for model training, the validation set is used for parameter adjustment and model optimization, and the test set is used for final model evaluation.

[0071] Step S14: Perform data processing, feature extraction, and time series modeling on the logging curve data to be processed in turn based on the target logging sedimentary microfacies identification model to determine the sedimentary microfacies identification result corresponding to the logging curve data to be processed.

[0072] It should be understood that in this embodiment, after obtaining the trained target well logging sedimentary microfacies identification model, sedimentary microfacies identification operations can be performed based on this model. Specifically, first, after completing the model evaluation operation of the target well logging sedimentary microfacies identification model based on the target sedimentary microfacies identification test set and the model evaluation result meets the preset conditions, the target well logging sedimentary microfacies identification model receives the well logging curve data to be processed; then, the target well logging sedimentary microfacies identification model uses the CKN network and the preset outlier detection strategy to perform outlier detection on the well logging curve data to be processed to determine the second outlier detection result; the CKN network is a network obtained by introducing the KAN network architecture into the CNN network to replace the original activation function; then, when the second outlier detection result indicates that there are outliers in the well logging curve data to be processed, the corresponding second outlier processing operation is triggered through the CKN network and based on the preset outlier processing strategy; after completing the second outlier processing operation, the CKN network and multiple pre-configured correlation coefficients are used to perform correlation analysis on the well logging curve data to be processed to determine the correlation analysis result; the correlation analysis result is displayed through the CKN network and the preset result display strategy to complete the well logging curve screening operation and determine the second screening result; the second screening result is standardized through the CKN network and based on the preset standard layer feature information to complete the data processing operation corresponding to the well logging curve data to be processed and determine the processed target well logging curve data; the CKN network is used to extract features from the target well logging curve data to determine the feature extraction result; time series modeling is performed through the LSTM network in the target well logging sedimentary microfacies identification model and the feature extraction result to determine the sedimentary microfacies identification result corresponding to the well logging curve data to be processed. As Figure 7 shown, the predicted category in the sedimentary microfacies identification result can also be compared with the true category.

[0073] Regarding outlier processing, in this embodiment, the outliers in the well logging curve data to be processed can be deleted based on the second outlier detection result; or, the mean of the first target data segment corresponding to the outliers in the well logging curve data to be processed is determined through the mean generation rule, and the outliers are replaced based on the mean; or, the median of the second target data segment corresponding to the outliers in the well logging curve data to be processed is determined through the median determination rule, and the outliers are replaced based on the median. In this way, the integrity and rationality of the data can be maintained.

[0074] Meanwhile, the logging data is standardized based on the standard layer features to eliminate the influence of non-geological factors caused by measurement system differences, ensuring the comparability and consistency of data in different well areas. A standard layer with stable geological features is selected as a reference, and the frequency distribution histogram method is used for analysis and correction to make the logging response features to be standardized consistent with the standard layer. That is, first, the target standard layer is selected from the second screening result through the CKN network and based on the preset standard layer feature information; then, the frequency distribution histogram of the curves of all single wells in the whole area at the target standard layer is drawn through the target standard layer and the second screening result, and data analysis and correction are carried out based on the frequency distribution histogram to complete the standardization operation.

[0075] In summary, regarding the model architecture of the target sedimentary microfacies recognition model, by replacing the traditional fixed activation function with KAN in the CNN structure, it can adaptively learn the non-linear features in the logging data, and combined with the time series modeling ability of LSTM, it effectively enhances the spatial continuity of sedimentary microfacies recognition and enables the model to more accurately depict the distribution characteristics of sedimentary facies belts. In addition, KAN can provide learnable geological information constraints during the multi-layer feature extraction process of CNN, improving the geological rationality of the model.

[0076] Furthermore, when using the target sedimentary microfacies recognition model to identify logging curve data, the CKN (CNN-KAN) structure is used to directly process the original logging curve data, rather than relying on image conversion, avoiding information loss caused by the limitation of the fixed activation function in the traditional CNN feature extraction process. As a learnable activation function of CNN, KAN makes the feature extraction process more adaptive and improves the expression ability of complex sedimentary microfacies features. And, compared with the traditional BiLSTM (Bidirectional Long Short-Term Memory), this method uses the CKN-LSTM structure. After the deep feature extraction of logging data is completed in the CNN-KAN stage, the time series relationship is modeled through LSTM, reducing the redundancy of the BiLSTM bidirectional calculation and improving the calculation efficiency. At the same time, the learnable activation mechanism of KAN reduces the need for manual adjustment of model parameters, making the training more efficient, with a faster convergence speed, and applicable to large-scale data sets. In addition, KAN in the CKN-LSTM structure can adaptively adjust the multi-layer features extracted by CNN, enabling the model to more effectively distinguish key sedimentary features from noise and improving the robustness of the model. In addition, LSTM further models the temporal relationship of logging data, making the recognition effect of the model more stable and the generalization ability stronger on different logging data sets in different regions.

[0077] That is to say, the sedimentary microfacies identification technical solution proposed in this embodiment has the following beneficial effects:

[0078] 1). By directly processing the original logging curve data with CKN, the information loss caused by image conversion is avoided. At the same time, combined with the adaptive characteristics of KAN, the feature extraction is more accurate, enabling the model to make full use of the complete information of the logging data, and improving the accuracy and reliability of classification;

[0079] 2). Using KAN as the activation module of CNN can adaptively adjust the non-linear mapping according to the distribution of logging data in different sedimentary environments, improve the feature expression ability, enhance the geological rationality of the model, enhance the recognition ability of complex formation microfacies, and thus significantly improve the classification accuracy;

[0080] 3). On the basis of CKN extracting spatial features, an LSTM network is further introduced. By using its powerful time series modeling ability, the historical evolution law of sedimentary microfacies can be effectively captured, the description ability of complex formation structures can be improved, so as to realize the fusion of time series information, enhance the time series consistency of sedimentary microfacies identification, and improve the stability of microfacies identification.

[0081] It can be seen that in the CNN network of this application, the original activation function is replaced with the KAN network, and an initial logging sedimentary microfacies identification model with logging curve data as the input is constructed in combination with the LSTM network. Then, the collected logging curve data and sedimentary microfacies interpretation data are processed to determine the target sedimentary microfacies identification training set. Then, the initial logging sedimentary microfacies identification model is trained using the target sedimentary microfacies identification training set. After the training is completed, the data processing, feature extraction and time series modeling are directly carried out on the to-be-processed logging curve data based on the determined target logging sedimentary microfacies identification model to determine the sedimentary microfacies identification result. In this way, the anti-noise ability, feature extraction ability and generalization performance of the logging sedimentary microfacies identification model can be effectively improved, the accuracy and precision of identification can be improved, the adaptability to complex sedimentary environments can be enhanced, and the computational complexity can be reduced.

[0082] See Figure 8 As shown, this application embodiment also correspondingly discloses a logging sedimentary microfacies identification device based on CKN-LSTM, including:

[0083] A data collection module 11, configured to collect logging curve data and sedimentary microfacies interpretation data corresponding to the logging curve data;

[0084] A sample construction module 12, configured to process the logging curve data based on a preset data processing rule, and construct samples based on the determined processed logging curve data and the sedimentary microfacies interpretation data to determine a target sedimentary microfacies identification training set and a target sedimentary microfacies identification verification set;

[0085] The model training module 13 is configured to build an initial logging sedimentary microfacies identification model with logging curve data as the input based on CKN-LSTM, and use the target sedimentary microfacies identification training set and the target sedimentary microfacies identification validation set to perform model training and iterative optimization on the initial logging sedimentary microfacies identification model until the model training operation is completed, and then determine the target logging sedimentary microfacies identification model; the CKN-LSTM is a neural network constructed by introducing the KAN network architecture into the CNN network to replace the original activation function and combining with the LSTM network.

[0086] The sedimentary microfacies identification module 14 is configured to perform data processing, feature extraction, and time series modeling on the logging curve data to be processed in sequence based on the target logging sedimentary microfacies identification model, so as to determine the sedimentary microfacies identification result corresponding to the logging curve data to be processed.

[0087] Among them, for the more specific working processes of the above-mentioned various modules, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details will not be elaborated herein.

[0088] It can be seen that in the CNN network of the present application, the original activation function is replaced with the KAN network, and an initial logging sedimentary microfacies identification model with logging curve data as the input is constructed by combining with the LSTM network. Then, data processing is performed on the collected logging curve data and sedimentary microfacies interpretation data to determine the target sedimentary microfacies identification training set. Then, the initial logging sedimentary microfacies identification model is trained using the target sedimentary microfacies identification training set. After the training is completed, data processing, feature extraction, and time series modeling are directly performed on the logging curve data to be processed in sequence based on the determined target logging sedimentary microfacies identification model to determine the sedimentary microfacies identification result. In this way, the anti-noise ability, feature extraction ability, and generalization performance of the logging sedimentary microfacies identification model can be effectively improved, the accuracy and precision of the identification can be improved, the adaptability to complex sedimentary environments can be enhanced, and the computational complexity can be reduced.

[0089] In some specific embodiments, the data collection module 11 may specifically include:

[0090] The first data collection unit is configured to collect logging curve data based on a first preset data source;

[0091] The second data collection unit is configured to collect sedimentary microfacies interpretation data corresponding to the logging curve data based on a second preset data source; the sedimentary microfacies interpretation data includes core analysis data.

[0092] In some specific embodiments, the sample construction module 12 may specifically include:

[0093] The first outlier detection unit is configured to perform outlier detection on the logging curve data based on a preset outlier detection strategy to determine the first outlier detection result;

[0094] The first outlier processing unit is configured to, when the first outlier detection result indicates that there are outliers in the logging curve data, trigger a corresponding first outlier processing operation based on a preset outlier processing strategy;

[0095] The first curve screening unit is configured to, after completing the first outlier processing operation, determine the target correlation coefficient value between different logging curves by analyzing the logging curve data, and perform logging curve screening based on the target correlation coefficient value to obtain the first screening result;

[0096] The first normalization processing unit is configured to perform normalization processing on the logging curve data in the first screening result based on preset standard layer feature information to determine the processed logging curve data.

[0097] In some specific embodiments, the sample construction module 12 may specifically include:

[0098] The label generation unit is configured to generate sedimentary microfacies labels corresponding to the processed logging curve data based on the sedimentary microfacies interpretation data and a preset expert annotation interface;

[0099] The data annotation unit is configured to annotate the processed logging curve data based on the sedimentary microfacies labels and a preset annotation rule to determine the target samples;

[0100] The sample division unit is configured to divide the target samples based on a preset data ratio to determine the target sedimentary microfacies recognition training set, the target sedimentary microfacies recognition verification set, and the target sedimentary microfacies recognition test set.

[0101] In some specific embodiments, the sedimentary microfacies recognition module 14 may specifically include:

[0102] The data receiving unit is configured to receive the logging curve data to be processed through the target logging sedimentary microfacies recognition model after completing the model evaluation operation of the target logging sedimentary microfacies recognition model based on the target sedimentary microfacies recognition test set and the model evaluation result meets the preset conditions;

[0103] The second outlier detection unit is configured to perform outlier detection on the logging curve data to be processed through the CKN network in the target logging sedimentary microfacies recognition model and the preset outlier detection strategy to determine the second outlier detection result; the CKN network is a network obtained by introducing the KAN network architecture into the CNN network to replace the original activation function;

[0104] A second outlier processing unit, configured to, when the second outlier detection result indicates that there are outliers in the well logging curve data to be processed, trigger a corresponding second outlier processing operation through the CKN network and based on the preset outlier processing strategy;

[0105] A correlation analysis unit, configured to, after completing the second outlier processing operation, perform correlation analysis on the well logging curve data to be processed respectively through the CKN network and a plurality of pre-configured correlation coefficients, so as to determine a correlation analysis result;

[0106] A second curve screening unit, configured to display the correlation analysis result through the CKN network and a preset result plotting and display strategy, so as to complete the well logging curve screening operation and determine a second screening result;

[0107] A second normalization processing unit, configured to perform normalization processing on the second screening result through the CKN network and based on preset standard layer feature information, so as to complete the data processing operation corresponding to the well logging curve data to be processed and determine the processed target well logging curve data;

[0108] A feature extraction unit, configured to extract features from the target well logging curve data through the CKN network, so as to determine a feature extraction result;

[0109] A time series modeling unit, configured to perform time series modeling through the LSTM network in the target well logging sedimentary microfacies identification model and the feature extraction result, so as to determine a sedimentary microfacies identification result corresponding to the well logging curve data to be processed.

[0110] In some specific embodiments, the second normalization processing unit is configured to specifically include:

[0111] A standard layer selection subunit, configured to select a standard layer from the second screening result through the CKN network and based on preset standard layer feature information, so as to determine a target standard layer;

[0112] A data analysis subunit, configured to draw a frequency distribution histogram of the curves of all wells in the whole area in the target standard layer through the target standard layer and the second screening result, and perform data analysis and correction based on the frequency distribution histogram, so as to complete the normalization processing operation.

[0113] In some specific embodiments, the second outlier processing unit specifically includes:

[0114] An outlier deletion subunit, configured to delete outliers in the well logging curve data to be processed based on the second outlier detection result;

[0115] The first outlier replacement subunit is configured to, alternatively, determine the mean of the first target data segment corresponding to the outlier in the to-be-processed logging curve data through a mean generation rule, so as to perform outlier replacement based on the mean.

[0116] The second outlier replacement subunit is configured to, alternatively, determine the median of the second target data segment corresponding to the outlier in the to-be-processed logging curve data through a median determination rule, so as to perform outlier replacement based on the median.

[0117] Furthermore, an embodiment of the present application also discloses an electronic device. Figure 9 It is a structural diagram of an electronic device 20 shown according to an exemplary embodiment, and the content in the figure should not be considered as any limitation on the scope of use of the present application.

[0118] Figure 9 This is a schematic structural diagram of an electronic device 20 provided by an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. Among them, the memory 22 is used to store a computer program, and the computer program is loaded and executed by the processor 21 to implement the relevant steps in the logging sedimentary microfacies identification method based on CKN-LSTM disclosed in any of the foregoing embodiments. In addition, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0119] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows is any communication protocol applicable to the technical solution of the present application, and no specific limitation is imposed on it here; the input / output interface 25 is used to obtain external input data or output data to the outside, and its specific interface type can be selected according to specific application needs, and no specific limitation is made here.

[0120] In addition, as a carrier for resource storage, the memory 22 may be a read-only memory, a random access memory, a disk, or an optical disc, etc., and the resources stored thereon may include an operating system 221, a computer program 222, etc., and the storage method may be short-term storage or permanent storage.

[0121] Among them, the operating system 221 is used to manage and control each hardware device and computer program 222 on the electronic device 20, and it can be Windows Server, Netware, Unix, Linux, etc. In addition to the computer program that can be used to complete the CKN-LSTM-based logging sedimentary microfacies identification method executed by the electronic device 20 disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs that can be used to complete other specific tasks.

[0122] Furthermore, the present application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the foregoing disclosed CKN-LSTM-based logging sedimentary microfacies identification method. For the specific steps of this method, reference may be made to the corresponding content disclosed in the foregoing embodiments, and details will not be repeated here.

[0123] In this specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0124] Those skilled in the art can further realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0125] The steps of the method or algorithm described in combination with the embodiments disclosed in this article can be directly implemented by hardware, a software module executed by a processor, or a combination of the two. The software module can be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, register, hard disk, removable disk, CD-ROM, or any other form of storage medium well-known in the technical field.

[0126] Finally, it should also be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising said element.

[0127] The technical solutions provided in this application have been introduced in detail above. Specific examples are used in this text to elaborate on the principles and implementation manners of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application; at the same time, for those of ordinary skill in the art, according to the idea of this application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. A logging sedimentary microfacies identification method based on CKN-LSTM, characterized in that, Including: Collecting logging curve data and sedimentary microfacies interpretation data corresponding to the logging curve data; Processing the logging curve data based on a preset data processing rule, and constructing samples based on the determined processed logging curve data and the sedimentary microfacies interpretation data to determine a target sedimentary microfacies recognition training set and a target sedimentary microfacies recognition verification set; Constructing an initial logging sedimentary microfacies recognition model with the input of logging curve data based on CKN-LSTM, and training and iteratively optimizing the initial logging sedimentary microfacies recognition model using the target sedimentary microfacies recognition training set and the target sedimentary microfacies recognition verification set until the model training operation is completed, and determining a target logging sedimentary microfacies recognition model; the CKN-LSTM is a neural network constructed by introducing a KAN network architecture into a CNN network to replace the original activation function and combining with an LSTM network; Based on the target logging sedimentary microfacies recognition model, performing data processing, feature extraction, and time series modeling on the logging curve data to be processed in sequence to determine a sedimentary microfacies recognition result corresponding to the logging curve data to be processed.

2. The logging sedimentary microfacies identification method based on CKN-LSTM according to claim 1, wherein The collecting of the logging curve data and the sedimentary microfacies interpretation data corresponding to the logging curve data includes: Collecting logging curve data based on a first preset data source; Collecting sedimentary microfacies interpretation data corresponding to the logging curve data based on a second preset data source; the sedimentary microfacies interpretation data includes core analysis data.

3. The logging sedimentary microfacies identification method based on CKN-LSTM according to claim 1, characterized in that The processing of the logging curve data based on the preset data processing rule includes: Performing outlier detection on the logging curve data based on a preset outlier detection strategy to determine a first outlier detection result; When the first outlier detection result indicates that there are outliers in the logging curve data, triggering a corresponding first outlier processing operation based on a preset outlier processing strategy; After completing the first outlier processing operation, determining a target correlation coefficient value between different logging curves by analyzing the logging curve data, and performing logging curve screening based on the target correlation coefficient value to obtain a first screening result; Performing standardization processing on the logging curve data in the first screening result based on preset standard layer feature information to determine the processed logging curve data.

4. The logging sedimentary microfacies identification method based on CKN-LSTM according to claim 3, wherein The constructing of samples based on the determined processed logging curve data and the sedimentary microfacies interpretation data to determine a target sedimentary microfacies recognition training set and a target sedimentary microfacies recognition verification set includes: Generating sedimentary microfacies labels corresponding to the processed logging curve data based on the sedimentary microfacies interpretation data and a preset expert annotation interface; Annotating the processed logging curve data based on the sedimentary microfacies labels and a preset annotation rule to determine target samples; Dividing the target samples based on a preset data ratio to determine a target sedimentary microfacies recognition training set, a target sedimentary microfacies recognition verification set, and a target sedimentary microfacies recognition test set.

5. The logging sedimentary microfacies identification method based on CKN-LSTM according to claim 4, wherein The performing of data processing, feature extraction, and time series modeling on the logging curve data to be processed in sequence based on the target logging sedimentary microfacies recognition model includes: After completing the model evaluation operation of the target log sedimentary microfacies recognition model based on the target sedimentary microfacies recognition test set and the model evaluation result meets the preset conditions, the target log sedimentary microfacies recognition model receives the log curve data to be processed; The CKN network in the target log sedimentary microfacies recognition model and the preset outlier detection strategy are used to detect outliers in the log curve data to be processed to determine the second outlier detection result; the CKN network is a network obtained by introducing the KAN network architecture into the CNN network to replace the original activation function; When the second outlier detection result indicates that there are outliers in the log curve data to be processed, the corresponding second outlier processing operation is triggered through the CKN network and based on the preset outlier processing strategy; After completing the second outlier processing operation, the CKN network and multiple preconfigured correlation coefficients are used to perform correlation analysis on the log curve data to be processed to determine the correlation analysis result; The CKN network and the preset result plotting and display strategy are used to display the correlation analysis result to complete the log curve screening operation and determine the second screening result; The CKN network and based on the preset standard layer feature information are used to perform normalization processing on the second screening result to complete the data processing operation corresponding to the log curve data to be processed and determine the processed target log curve data; The CKN network is used to extract features from the target log curve data to determine the feature extraction result; The LSTM network in the target log sedimentary microfacies recognition model and the feature extraction result are used to perform time series modeling to determine the sedimentary microfacies recognition result corresponding to the log curve data to be processed.

6. The well logging sedimentary microfacies identification method based on CKN-LSTM according to claim 5, wherein, The performing normalization processing on the second screening result through the CKN network and based on the preset standard layer feature information includes: The CKN network and based on the preset standard layer feature information are used to select the standard layer from the second screening result to determine the target standard layer; The frequency distribution histogram of the curves of the whole area single well at the target standard layer is drawn through the target standard layer and the second screening result, and data analysis and correction are performed based on the frequency distribution histogram to complete the normalization processing operation.

7. The logging sedimentary microfacies identification method based on CKN-LSTM according to claim 5, wherein The triggering the corresponding second outlier processing operation through the CKN network and based on the preset outlier processing strategy includes: Deleting the outliers in the log curve data to be processed based on the second outlier detection result; Or, determining the mean of the first target data segment corresponding to the outliers in the log curve data to be processed through the mean generation rule to perform outlier replacement based on the mean; Or, determining the median of the second target data segment corresponding to the outliers in the log curve data to be processed through the median determination rule to perform outlier replacement based on the median.

8. A logging sedimentary microfacies identification device based on CKN-LSTM, characterized in that, including: The data collection module is used to collect log curve data and the sedimentary microfacies interpretation data corresponding to the log curve data; A sample construction module, configured to process the logging curve data based on a preset data processing rule, and construct a sample based on the determined processed logging curve data and the sedimentary microfacies interpretation data, so as to determine a target sedimentary microfacies identification training set and a target sedimentary microfacies identification verification set; A model training module, configured to construct an initial logging sedimentary microfacies identification model with the logging curve data as the input based on CKN-LSTM, and use the target sedimentary microfacies identification training set and the target sedimentary microfacies identification verification set to perform model training and iterative optimization on the initial logging sedimentary microfacies identification model, until when the model training operation is completed, determine a target logging sedimentary microfacies identification model; the CKN-LSTM is a neural network constructed by introducing a KAN network architecture into a CNN network to replace the original activation function and combining with an LSTM network; A sedimentary microfacies identification module, configured to sequentially perform data processing, feature extraction, and time series modeling on the logging curve data to be processed based on the target logging sedimentary microfacies identification model, so as to determine a sedimentary microfacies identification result corresponding to the logging curve data to be processed.

9. An electronic device, characterized in that, Comprising: A memory, configured to store a computer program; A processor, configured to execute the computer program to implement the CKN-LSTM-based logging sedimentary microfacies identification method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, For storing a computer program, the computer program, when executed by a processor, implements the CKN-LSTM-based logging sedimentary microfacies identification method according to any one of claims 1 to 7.

Images