Transverse wave velocity prediction method, device and equipment

By screening target logging parameters for each known well and establishing a decision tree model, cross-wave velocity prediction of the predicted well, and weighted processing with inter-well distance information, the low cross-wave velocity prediction accuracy and efficiency in the existing technology are solved. Problem, achieving higher prediction accuracy and efficiency.

CN118897322AActive Publication Date: 2024-11-05CHINA UNIV OF PETROLEUM (BEIJING)

Patent Information

Application Number
CN202410946029.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-15
Publication Date
2024-11-05
Estimated Expiration
2044-07-15

AI Technical Summary

Technical Problem

The prediction method of obtaining transverse wave velocity in the prior art is difficult to ensure prediction accuracy and prediction efficiency. It is mainly because the geological environment differences between each known well are ignored, resulting in the model being unable to accurately capture the characteristics of each well, which in turn affects the prediction of transverse wave velocity. Accuracy.

Method used

By screening out the target logging parameters based on its own logging parameters and transverse wave velocity for each known well, and establishing a decision tree model, using these models to predict transverse wave velocity for the predicted well, taking into account the distance information between the wells Weighted processing to improve prediction accuracy and efficiency.

Benefits of technology

This method can capture the characteristics of each well more accurately, reduce prediction errors, improve the applicability and robustness of the model in different environments, and significantly improve the accuracy and efficiency of cross-wave velocity prediction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118897322A_ABST
    Figure CN118897322A_ABST
Patent Text Reader

Abstract

The invention provides a shear wave velocity prediction method, device and equipment, and the method comprises the steps: determining a plurality of target logging parameters from all logging parameters for each known well according to a first data set of the known well, the influence degree of the target logging parameter on the shear wave velocity of the known well is greater than that of other logging parameters on the shear wave velocity of the known well; according to the second data set of the known well, constructing a decision tree model of the known well; a to-be-predicted data set of the to-be-predicted well is input into the decision tree model of the known well, the initial shear wave velocity output by the decision tree model is obtained, and the to-be-predicted data set comprises the parameter value of each target well logging parameter of the to-be-predicted well under each preset depth; and generating the target shear wave velocity of the well to be predicted according to the initial shear wave velocity output by the known well in each decision tree model. According to the method provided by the invention, the shear wave velocity prediction precision and prediction efficiency are effectively improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of oil and gas exploration, and in particular to a shear wave velocity prediction method, device and equipment. Background Art

[0002] In the exploration and development of oil and gas, P-wave and S-wave velocities provide key underground information, and S-wave velocity is a key parameter in rock physics research, which is conducive to the identification of pore fluids and reservoir properties, and plays a very important role in prestack inversion and Amplitude Variation with Offset (AVO) attribute analysis. The inversion of combined P-wave and S-wave velocities helps to reduce the multi-solution of the reservoir. Therefore, obtaining S-wave velocity data is important for geophysical exploration.

[0003] The existing methods for obtaining shear wave velocity are usually based on prediction methods, which mainly use data-driven machine learning to predict shear wave velocity data. Specifically, the shear wave velocity data of multiple existing wells and the corresponding logging parameters (such as acoustic time difference, density, natural gamma, etc.) are first collected for data preprocessing, and then the nonlinear mapping relationship is learned and processed by machine learning to complete the model training to predict the shear wave velocity of the well to be predicted.

[0004] However, the shear wave velocity obtained by the prediction method in the prior art is difficult to ensure the prediction accuracy and prediction efficiency. Summary of the invention

[0005] The present application provides a shear wave velocity prediction method, device and equipment to solve the problem that the shear wave velocity obtained by the prediction method in the prior art is difficult to ensure the prediction accuracy and prediction efficiency.

[0006] In a first aspect, the present application provides a method for predicting shear wave velocity, comprising:

[0007] For each known well, a plurality of target logging parameters are determined from all logging parameters according to a first data set of the known well, wherein the first data set includes the shear wave velocity of the known well at each preset depth and the parameter value of each logging parameter, and the influence of the target logging parameter on the shear wave velocity of the known well is greater than the influence of other logging parameters on the shear wave velocity of the known well;

[0008] constructing a decision tree model of the known well according to a second data set of the known well, wherein the second data set includes the shear wave velocity of the known well at each preset depth and the parameter value of each target logging parameter;

[0009] Inputting a data set to be predicted of the well to be predicted into the decision tree model of the known well to obtain an initial shear wave velocity output by the decision tree model, wherein the data set to be predicted includes a parameter value of each target logging parameter of the well to be predicted at each preset depth;

[0010] According to the initial shear wave velocity output by the known well in each decision tree model, the target shear wave velocity of the well to be predicted is generated.

[0011] In a second aspect, the present application provides a shear wave velocity prediction device, comprising:

[0012] a determination module, configured to determine, for each known well, a plurality of target logging parameters from all logging parameters according to a first data set of the known well, wherein the first data set includes the shear wave velocity of the known well at each preset depth and a parameter value of each logging parameter, wherein the target logging parameter has a greater influence on the shear wave velocity of the known well than other logging parameters;

[0013] A construction module, configured to construct a decision tree model of the known well according to a second data set of the known well, wherein the second data set includes a shear wave velocity of the known well at each preset depth and a parameter value of each target logging parameter;

[0014] An input module, used for inputting a data set to be predicted of the well to be predicted into the decision tree model of the known well, and obtaining an initial shear wave velocity output by the decision tree model, wherein the data set to be predicted includes a parameter value of each target logging parameter of the well to be predicted at each preset depth;

[0015] The generating module is used to generate the target shear wave velocity of the well to be predicted according to the initial shear wave velocity output by the known well in each decision tree model.

[0016] Optionally, the determining module is specifically used to:

[0017] For each well logging parameter, extracting a third data set corresponding to the well logging parameter from the first data set of the known well, the third data set including the shear wave velocity of the known well at each preset depth and the parameter value of the well logging parameter;

[0018] For each parameter value of the logging parameter, splitting a third data set corresponding to the logging parameter according to the parameter value to generate a first sub-data set and a second sub-data set;

[0019] Calculate, according to the first sub-data set, the second sub-data set and a third data set corresponding to the logging parameter, a mean square error (MSE) reduction value of the logging parameter before and after splitting according to the parameter value;

[0020] Calculating the MSE cumulative reduction value of the logging parameter according to the MSE reduction value corresponding to each parameter value of the logging parameter;

[0021] According to the MSE cumulative reduction value of each logging parameter, a plurality of target logging parameters are determined from all the logging parameters.

[0022] In a third aspect, the present application provides an electronic device, comprising: a processor, and a memory communicatively connected to the processor;

[0023] Memory stores computer-executable instructions;

[0024] The processor executes the computer-executable instructions stored in the memory to implement the shear wave velocity prediction method as described in the first aspect and various possible implementations of the first aspect.

[0025] The shear wave velocity prediction method, device and equipment provided by the present application, for each known well, according to the first data set of the known well, multiple target logging parameters are determined from all logging parameters, the first data set includes the shear wave velocity of the known well at each preset depth and the parameter value of each logging parameter, and the influence of the target logging parameter on the shear wave velocity of the known well is greater than the influence of other logging parameters on the shear wave velocity of the known well; according to the second data set of the known well, a decision tree model of the known well is constructed; the data set to be predicted of the well to be predicted is input into the decision tree model of the known well to obtain the initial shear wave velocity output by the decision tree model; according to the initial shear wave velocity output by the known well in each decision tree model, the target shear wave velocity of the well to be predicted is generated. This method utilizes the characteristics that the logging parameters of each known well are different due to different geological environments, and starts with independently screening the target logging parameters of each known well, establishing a decision tree model for each known well, and finally generating a decision tree model. When the shear wave velocity of the well to be predicted needs to be predicted, the model can more accurately capture the characteristics of each known well, thereby reducing the prediction error caused by using the common logging parameters of multiple wells. At the same time, it improves the applicability and robustness of the model in different environments and improves the model prediction efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0027] Figure 1 Schematic diagram of the process of the shear wave velocity prediction method provided in the embodiment of the present application Figure 1 ;

[0028] Figure 2 A schematic diagram of the coordinates of each known well in the park and the corresponding distance weight provided in the embodiment of the present application;

[0029] Figure 3 The intersection diagram of the predicted values ​​and true values ​​of the four different machine learning methods for the wells to be predicted provided in the embodiments of the present application;

[0030] Figure 4 A curve comparison diagram of the actual longitudinal and transverse wave velocity ratio provided in the embodiment of the present application and the longitudinal and transverse wave velocity ratio predicted by different methods;

[0031] FIG5( a ) is a comparison diagram of shear wave velocity prediction curves under low-pass filtering using different prediction methods provided in an embodiment of the present application;

[0032] FIG5( b ) is a histogram comparison of shear wave velocity prediction errors of different prediction methods provided in an embodiment of the present application under low-pass filtering;

[0033] Figure 6 Schematic diagram of the process of the shear wave velocity prediction method provided in the embodiment of the present application Figure 2 ;

[0034] Figure 7(a) shows the importance score ranking of the known well A1 logging parameters;

[0035] Figure 7(b) shows the ranking of the importance scores of the logging parameters of the known well A2;

[0036] Figure 7(c) shows the ranking of the importance scores of the logging parameters of the known well A3;

[0037] Figure 8 A linear correlation thermodynamic diagram between each well logging parameter of the well to be predicted provided in an embodiment of the present application;

[0038] Fig. 9 A schematic diagram of an interface of a shear wave velocity prediction method provided in an embodiment of the present application;

[0039] Fig.10 A schematic diagram of the structure of a shear wave velocity prediction device provided in an embodiment of the present application;

[0040] Fig.11 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0041] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0042] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0043] 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 laws, regulations and standards, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0044] In the exploration and development of oil and gas, P-wave and S-wave velocities provide key underground information. S-wave velocity is helpful in identifying pore fluids and reservoir properties, and plays a very important role in prestack inversion and AVO attribute analysis. The inversion of combined P-wave and S-wave velocities helps reduce the multi-solution of the reservoir. Among them, S-wave velocity can be obtained through laboratory core analysis or dipole acoustic logging. Although these logging methods can provide high-quality data, they are expensive to implement and have low measurement efficiency under complex geological conditions. In addition, for old wells that have been drilled, there are usually no conditions for these special logging, which leads to the loss of S-wave velocity data and limits the ability to describe and analyze the reservoir in detail. Therefore, the use of prediction to obtain S-wave velocity data is widely used by researchers and engineers.

[0045] At present, the prediction method of shear wave velocity is usually mainly based on data-driven machine learning to predict shear wave velocity data. Specifically, by first collecting the shear wave velocity data of multiple existing wells and the corresponding logging parameters (such as acoustic time difference, density, natural gamma, etc.), data preprocessing (including denoising, missing value filling and standardization) is carried out to ensure the quality and consistency of the data. Then, machine learning is used to learn and process the nonlinear mapping relationship, complete the model training and apply it to the logging data of the wells to be predicted in the unmeasured area to predict the shear wave velocity.

[0046] However, in the prior art, most machine learning methods usually only use multiple logging parameters and corresponding shear wave velocities of multiple known wells, and screen out the common target logging parameters of multiple known wells in units of multiple wells. Since the differences between the known wells are ignored, errors are easily introduced during modeling. For example, the geological environment of different known wells may cause significant differences in logging parameters, and the use of public target logging parameters will cause the model to be unable to accurately capture the characteristics of each known well, thereby affecting the prediction accuracy of shear wave velocity. Secondly, this method is also difficult to guarantee the improvement of prediction efficiency. Due to the failure to fully utilize the data characteristics of each known well, the model may require more adjustments and optimizations in actual applications, which increases the computing cost and time consumption.

[0047] Based on this, the present application proposes a shear wave velocity prediction method, device and equipment. Due to the differences in the geological environment of the known wells, the logging parameters of each known well may be significantly different. The target logging parameters can be targeted for different known wells, that is, taking a known well as a unit, the target logging parameters corresponding to each known well are screened according to its own logging parameters and the corresponding shear wave velocity. Further, a decision tree model is generated using the target logging parameters corresponding to each known well. When it is necessary to predict the well to be predicted, the decision tree model is used to output the predicted shear wave velocity of each known well, and then the target shear wave velocity of the well to be predicted is obtained. This method independently screens and analyzes the target logging parameters and establishes a decision tree model according to the unique logging parameters of each known well, so that the decision tree model can capture the target logging parameters of each well more accurately, thereby improving the accuracy of the prediction, and selecting the target logging parameters in units of a well can make the decision tree model more concise and targeted, thereby reducing the amount of calculation and improving the operating efficiency of the decision tree model.

[0048] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0049] Figure 1 Schematic diagram of the process of the shear wave velocity prediction method provided in the embodiment of the present application Figure 1 .like Figure 1 As shown, the method includes:

[0050] S101, for each known well, determining a plurality of target logging parameters from all logging parameters according to a first data set of the known well;

[0051] The first data set includes the shear wave velocity of the known well at each preset depth and the parameter value of each logging parameter, and the target logging parameter has a greater impact on the shear wave velocity of the known well than other logging parameters.

[0052] The known well is a well for which the shear wave velocity at each preset depth and the parameter value of each logging parameter have been obtained in a certain area.

[0053] For example, it is assumed that the impact of the input G logging parameters on the shear wave velocity is sorted from high to low, and the first g logging parameters with higher impact are selected as the target logging parameters for finally constructing the decision tree model, and the selection of the target logging parameters is completed, that is, After filtering, it is transformed into

[0054] It can be understood that for each known well, a first data set is first constructed for the shear wave velocity of the known well at each preset depth and the parameter value of each logging parameter, and then the first data set is used to analyze the influence of all logging parameters of the known well on the shear wave velocity, and multiple logging parameters with high influence are selected as target logging parameters to construct a decision tree model, thereby eliminating redundant and irrelevant logging parameters, reducing the influence of noise on the decision tree model, making the decision tree model more stable, and enhancing the robustness and generalization ability of the decision tree model. In addition, the decision tree model can be enabled to more accurately capture the changing law of the shear wave velocity and reduce the prediction error, thus laying a foundation for improving the prediction accuracy.

[0055] S102, constructing a decision tree model of the known wells according to a second data set of the known wells;

[0056] The second data set includes the shear wave velocity of the known well at each preset depth and the parameter value of each target logging parameter.

[0057] Among them, the decision tree model is a weak classifier, which can output the predicted initial shear wave velocity of the well to be predicted based on the data set of the well to be predicted. The data set to be predicted includes the parameter value of each target logging parameter of the well to be predicted at each preset depth.

[0058] It can be understood that by determining the parameter value of each target logging parameter of the known well and the shear wave velocity at each preset depth, a second data set can be established, and then a decision tree model can be established for each known well based on the data set, thereby constructing a complete decision tree model of the known wells for subsequent shear wave velocity prediction of the wells to be predicted.

[0059] The decision tree model is an inverted tree structure, including a root node and a leaf node. The root node contains the complete data set, that is, the second data set. The leaf node is the split node of the decision tree, that is, the node that splits the data set. In decision tree regression, the tree construction process is to recursively divide the data set into smaller and smaller subsets through the split point in a greedy algorithm. The greedy algorithm is an algorithm that selects the best choice in the current state at each step. Its goal is to eventually reach the global optimal solution through the local optimal solution.

[0060] In the decision tree splitting process of a known well, a logging parameter and the corresponding parameter value are selected as the root node, and then the data set is split into two similar data sets to generate the next node. The next node can continue to split until the stop condition is met (such as the maximum depth of the tree, the minimum number of samples in the node, etc.). The decision tree calculates the average value of the target shear wave velocity at each leaf node as the final prediction result of the area.

[0061] S103, inputting the to-be-predicted data set of the to-be-predicted well into the decision tree model of the known well, and obtaining the initial shear wave velocity output by the decision tree model;

[0062] The data set to be predicted includes the parameter value of each target logging parameter of the well to be predicted at each preset depth.

[0063] The initial shear wave velocity refers to the shear wave velocity of the well to be predicted outputted for each known well in the prediction model after the data set of the well to be predicted is input into the decision tree model of the known wells.

[0064] It can be understood that the data set to be predicted of the well to be predicted is input into the decision tree model composed of known wells for prediction. The decision tree models of all known wells are constructed into a random forest model, which can integrate the prediction results of multiple known wells and reduce the overfitting problem that may exist in the decision tree model of a single known well, thereby improving the accuracy of shear wave velocity prediction.

[0065] S104, generating a target shear wave velocity of the well to be predicted according to the initial shear wave velocity output by the known well in each decision tree model.

[0066] The target shear wave velocity refers to the final shear wave velocity of the well to be predicted, which is generated by processing the initial shear wave velocity output by each known well.

[0067] In a possible implementation, the initial shear wave velocity can be weighted to generate the final shear wave velocity of the well to be predicted. However, the commonly used weighting method fails to take into account the spatial similarity between the two wells and cannot assign appropriate weight values ​​to different initial shear wave velocities, which may lead to prediction errors due to differences in geological sedimentary laws between the well to be predicted and the known wells far away from the well to be predicted. Since the spatial similarity between the two wells is related to the distance between the two wells, the closer the distance between different wells, the more similar their geological sedimentary laws and logging parameter responses, and the higher the spatial similarity; conversely, the farther the distance, the lower the correlation of geological sedimentary laws, and the lower the spatial similarity. Therefore, the distance is further considered. When generating the target shear wave velocity of the well to be predicted based on the initial shear wave velocity output by the known wells in each decision tree model, the distance between each known well and the well to be predicted can also be considered as a weight value.

[0068] Figure 2 A schematic diagram of the coordinates of each known well in the park and the corresponding distance weight provided in the embodiment of the present application. Figure 2 (a) in the figure is used to indicate the position between different wells. There are three known wells A1, A2, and A3 in the park, and the well A4 to be predicted. The distances between wells A1, A2, and A3 and well A4 are arranged from near to far. Figure 2 (b) is the distance weight curve of each known well. According to the rule that the closer the distance, the higher the spatial similarity, it can be seen that the nearest well A1 has the largest weight value, followed by well A2, and finally well A3. When predicting the shear wave velocity of well A4, well A1, well A2 and well A3 give distance weights of 0.55, 0.27 and 0.18 respectively. Well A1 has the greatest impact on the prediction result of well A4.

[0069] In one achievable method, the location information of each known well and the location information of the well to be predicted can be first obtained, and then the target shear wave velocity of the well to be predicted is generated based on the initial shear wave velocity output by the decision tree model of the known well, the location information of the known well and the location information of the well to be predicted.

[0070] The location information refers to the geographical location coordinates corresponding to different wells. Such location information can be obtained through geographic information system databases, global positioning system data, and map data. In practical applications, the acquisition of location information can be determined according to actual conditions, and the embodiments of the present application do not impose specific restrictions on this.

[0071] It is understandable that, considering that the distance between each known well and the well to be predicted has a certain influence on the target shear wave velocity of the well to be predicted, the initial shear wave velocity output by each known well in the decision tree model is integrated by inverse distance weighting. Therefore, the location information of each known well and the well to be predicted is first obtained, and the initial shear wave velocity output by the decision tree model is weighted. According to the distance, different weight values ​​are assigned to the initial shear wave velocities output by different known wells, thereby obtaining the target shear wave velocity of the well to be predicted. This method effectively reduces the high error influence of the distant known wells on the prediction of the target shear wave velocity of the well to be predicted, and provides a more accurate prediction result.

[0072] In one achievable manner, the specific calculation formula for generating the target shear wave velocity of the well to be predicted is as follows based on the initial shear wave velocity output by the known well in each decision tree model, the location information of the known well, and the location information of the well to be predicted:

[0073]

[0074] Among them, y pred is the target shear wave velocity of the well to be predicted, is the initial shear wave velocity output by the decision tree model of the known well i, n is the number of known wells, b is the weighted power index, x is the horizontal coordinate of the well to be predicted, y is the vertical coordinate of the well to be predicted, and x i is the horizontal coordinate of the known well i, y i is the ordinate of the known well i. The abscissa and ordinate of the well to be predicted are determined according to the location information of the well to be predicted, and the abscissa and ordinate of the known well i are determined according to the location information of the known well i.

[0075] The weighted power index is a form of weighted average, which is a type of average defined by different power values ​​b. By selecting different power values ​​b, different types of data and application scenarios can be adapted, thereby providing a flexible tool for complex data analysis. The embodiment of the present application does not specifically limit the selection of the weighted power index b.

[0076] For example, suppose there are n known wells W 1 ,W 2 …W n , and the well to be predicted W p . When making predictions W p Before weighting the prediction results of different initial shear wave velocities, it is necessary to calculate the predicted W p With the known well W 1 ,W 2 …W n The distance between them. That is, using the well to be predicted W p The coordinates of the point (x, y), n known wells W 1 ,W2 …W n The coordinates of the point (x i ,y i ), i = 1, 2, 3 ... n. For each known well W 1 ,W 2 …W n and the well to be predicted W p There will be a reciprocal of the distance as the weight value, according to the weight values ​​of different known wells and the initial shear wave velocity output by different known wells in the decision tree model The product of is accumulated to obtain the target shear wave velocity y of the well to be predicted. pred .

[0077] It can be understood that by selecting the target logging parameters for each known well and establishing a decision tree model for each known well based on the target logging parameters, the prediction error caused by the low-impact logging parameters on the model is effectively reduced. And based on the initial shear wave velocity output by different known wells in the decision tree model, the inverse distance weighted integration strategy is adopted, and the distance information between different known wells and the well to be predicted is used as a weight to be assigned to the results of the initial shear wave velocity output by different known wells. Finally, the target shear wave velocity of the well to be predicted is determined, and a high-precision shear wave velocity prediction combining the similarity of position information between different wells is achieved, ensuring that the known wells closer to the well to be predicted have a greater weight on the prediction results, thereby improving the prediction accuracy and reliability of the model.

[0078] Figure 3 The intersection diagram of the predicted values ​​and true values ​​of the four different machine learning methods for the wells to be predicted provided in the embodiment of the present application; wherein the diamond represents the extreme gradient boosting (eXtreme Gradient Boosting, XGBoost), the square represents the random forest model (Random Forest, RF), the triangle represents the support vector regression model (Support Vector Regression, SVR) and the circle represents the geologically constrained random forest (Geologically constrained random forest, GCRF) model. It can be seen that the predicted values ​​of the four methods have a high correlation with the true value, and the determination coefficient (Rsquared, R 2 ) values ​​are all above 0.93. However, the GCRF method used in this application performs best, with the smallest mean absolute error (MAE) value, the most concentrated data point distribution, and the smallest prediction error. Therefore, the prediction accuracy of this model is the highest, and the stability of the model is good.

[0079] Figure 4The following is a curve comparison of the actual longitudinal and transverse wave velocity ratio provided in the embodiment of the present application and the longitudinal and transverse wave velocity ratio predicted by different methods. Figure 4 A in the figure is the comparison curve between the real P-wave velocity ratio and the P-wave velocity ratio predicted by the XGBoost method; Figure 4 B in the figure is the comparison curve between the real P-wave velocity ratio and the P-wave velocity ratio predicted by the RF method; Figure 4 C in the figure is the comparison curve between the real P-wave velocity ratio and the P-wave velocity ratio predicted by the SVR method; Figure 4 D in is the comparison curve between the real P-wave velocity ratio and the P-wave velocity ratio predicted by the GCRF method. Figure 4 It can be seen that at a depth of 1536m, the R of the actual P-wave velocity ratio obtained by the XGBoost method is different from the predicted P-wave velocity ratio. 2 The value is the smallest and the MAE value is the largest, indicating that the fit between the true P-wave velocity ratio and the predicted P-wave velocity ratio is poor and the prediction accuracy is low. 2 The MAE value is at a medium level, indicating that the prediction accuracy of the P-wave velocity ratio by these two models is better than that by the extreme gradient boosting method, and the R of the real P-wave velocity ratio obtained by the method shown in this application and the predicted P-wave velocity ratio is 2 The maximum and the MAE value are the smallest, indicating that the GCRF model constructed by selecting the target logging parameters for each well and establishing a decision tree model has a better fit between the predicted P-wave and S-wave velocity ratios and the true P-wave and S-wave velocity ratios, a smaller prediction error, and a higher prediction accuracy.

[0080] Further, FIG5(a) is a comparison diagram of the shear wave velocity prediction curves under low-pass filtering of different prediction methods provided in the embodiment of the present application. Among them, the black line represents the low-pass filtering result of the measured shear wave velocity, the gray triangle line represents the low-pass filtering result of the shear wave prediction obtained by the logistic regression (LR) method, and the gray square line represents the low-pass filtering result of the shear wave prediction obtained by the GCRF model. FIG5(b) is a comparison diagram of the shear wave velocity prediction error histogram under low-pass filtering of different prediction methods provided in the embodiment of the present application. Among them, the white squares represent the histogram of the shear wave velocity prediction error under low-pass filtering obtained by the LR method, and the gray squares represent the histogram of the shear wave velocity prediction error under low-pass filtering obtained by the LR method. It can be seen that the shear wave velocity prediction curve obtained by the GCRF model represented by the gray square line has a better fitting effect with the actual shear wave velocity curve, and the distribution of the shear wave velocity prediction error histogram under low-pass filtering obtained by the GCRF model is more stable, indicating that the method of the present application selects target logging parameters for each known well according to its own unique logging parameters, and establishes a decision tree model for each well separately, and finally constructs a complete random forest shear wave velocity prediction model. At the same time, considering the distance and spatial similarity between different wells, the distance is added, and the prediction results of the decision tree model of each known well are weighted, thereby improving the model prediction accuracy, accuracy and stability.

[0081] An embodiment of the present application provides a shear wave velocity prediction method, which determines a plurality of target logging parameters from all logging parameters for each known well according to a first data set of the known well, the first data set including the shear wave velocity of the known well at each preset depth and the parameter value of each logging parameter, and the influence of the target logging parameter on the shear wave velocity of the known well is greater than the influence of other logging parameters on the shear wave velocity of the known well; constructs a decision tree model of the known well according to a second data set of the known well; inputs the data set to be predicted of the well to be predicted into the decision tree model of the known well to obtain an initial shear wave velocity output by the decision tree model, the data set to be predicted including the parameter value of each target logging parameter at each preset depth of the well to be predicted; generates a target shear wave velocity of the well to be predicted according to the initial shear wave velocity output by the known well in each decision tree model. This method takes advantage of the fact that the logging parameters of each known well are different due to different geological environments. It starts with independently screening the target logging parameters of each known well, and establishes a complete decision tree model through the target logging parameters of each known well. Then, based on the initial shear wave velocities output by different known wells in the prediction model, an inverse distance weighted integration strategy is adopted to use the distance information between different known wells and the well to be predicted as weights and assign them to the results of the initial shear wave velocities output by different known wells. Finally, the target shear wave velocity of the well to be predicted is determined, ensuring that more accurate and reliable shear wave velocity prediction results can be provided under complex geological conditions and when the logging data between wells are quite different, thereby significantly improving the efficiency and accuracy of oil and gas exploration.

[0082] In one achievable manner, the specific process of constructing a decision tree model of a known well according to the second data set of the known well can be implemented by the following steps:

[0083] Step a, determining the well logging parameters to be processed from the target well logging parameters corresponding to the known wells;

[0084] The logging parameters to be processed refer to the logging parameters that need to be split subsequently.

[0085] It can be understood that after the target logging parameters corresponding to the known well are determined, the logging parameters to be processed for subsequent splitting are determined therefrom.

[0086] Step b, extracting a fourth data set corresponding to the well logging parameters to be processed from the second data set of the known well;

[0087] The fourth data set includes the shear wave velocity of the known well at each preset depth and the parameter values ​​of the well logging parameters to be processed;

[0088] It is understandable that after determining the logging parameters to be processed based on the target logging parameters of the known well, the shear wave velocity and parameter values ​​of the logging parameters to be processed corresponding to each preset depth of the known well are extracted from the second data set.

[0089] Exemplarily, Table 1 is a second data set of a known well, including the shear wave velocity at each preset depth and the parameter value of the target logging parameter. From the target logging parameter corresponding to the known well, it is determined that the logging parameter to be processed is a density parameter, and then the fourth data set corresponding to the logging parameter to be processed is extracted from the second data set as shown in Table 2, including the shear wave velocity of the known well at each preset depth and the parameter value of the logging parameter to be processed.

[0090] Table 1 The second data set of a known well

[0091] depth density Natural Gamma Shear wave velocity 1600 2.1 30 3.5 1650 2.8 32 4.0 1700 2.5 28 4.5 1750 3.0 29 5.0

[0092] Table 2 The fourth data set of a known well

[0093] depth density Shear wave velocity 1600 2.1 3.5 1650 2.8 4.0 1700 2.5 4.5 1750 3.0 5.0

[0094] Step c, according to the fourth data set, determining a target parameter value from the parameter values ​​of the logging parameters to be processed, and the total sum of squared errors (SSE) of the third sub-dataset and the fourth sub-dataset generated after splitting the fourth data set according to the target parameter value is less than the total SSE of the third sub-dataset and the fourth sub-dataset generated after splitting the fourth data set according to other parameter values;

[0095] The target parameter value refers to the value for splitting the fourth data set. The third sub-data set and the fourth sub-data set refer to two similar data sets formed after splitting the fourth data set according to the target parameter value.

[0096] For example, assuming that according to the fourth data set of known wells in Table 2, the density parameter is selected as 2.5 as the target parameter value, then the third sub-datasets generated after the fourth data set is split according to this value are [1600, 2.1, 3.5], [1700, 2.5, 4.5], and the fourth sub-datasets are [1650, 2.8, 4.0],

[0097] [1750, 3.0, 5.0], and the total sum of square errors SSE is calculated based on the third and fourth sub-datasets. The total sum of square errors SSE generated by splitting with a density parameter of 2.5 as the target parameter value is smaller than the total SSE of the third and fourth sub-datasets generated by splitting with density parameters of 2.1, 2.8, and 3.0 as parameter values.

[0098] It can be understood that by selecting the parameter value corresponding to the minimum total SSE as the target parameter value, it can be ensured that the generated shear wave velocity prediction error is the minimum in the area where the current target parameter value is split. This means that each split reduces the error in the sense of local optimality. And each split will divide the data more accurately, so that the model can better capture the characteristics of the parameters, thereby improving the prediction accuracy of the model.

[0099] In one achievable manner, according to the fourth data set, determining the target parameter value from the parameter values ​​of the well logging parameters to be processed includes:

[0100] First, for each parameter value of the logging parameter to be processed, the fourth data set is split according to the parameter value to generate the third sub-dataset and the fourth sub-dataset; then, the total SSE of the third sub-dataset and the fourth sub-dataset is calculated according to the third sub-dataset and the fourth sub-dataset; finally, the parameter value with the smallest total SSE is determined as the target parameter value.

[0101] Optionally, a specific formula for splitting the fourth data set according to the target parameter value is:

[0102] Q 1 (j,a)={X|X (j) ≤a},Q 2 (j,a)={X|X (j) >a}

[0103]

[0104] Among them, a is the target parameter value; j is the logging parameter to be processed; Q 1 is the third sub-dataset generated after splitting the fourth data set according to the target parameter value; Q 2 A fourth sub-dataset generated by splitting the fourth data set according to the target parameter value; According to the third sub-dataset Q 1 The corresponding output shear wave velocity or according to the fourth data set Q 2 Corresponding to the output shear wave velocity; N m Q 1 or Q 2 The number of shear wave velocities included in the data set; X is the shear wave velocity.

[0105] The specific formula for determining the target parameter value a from the parameter values ​​of the well logging parameters to be processed is:

[0106]

[0107] Among them, min is the minimum value; C 1 According to the third sub-dataset Q 1 Corresponding to the output shear wave velocity; C2 According to the fourth data set Q 2 The corresponding output shear wave velocity; is the SSE generated based on the third sub-dataset; is the SSE generated according to the fourth sub-dataset.

[0108] Continuing with the above example, assume that according to the fourth data set of known wells in Table 2, the density parameters are selected as 2.1, 2.5, 2.8, and 3.0 as parameter values, respectively. Then, the third and fourth sub-datasets generated after the fourth data set is split according to these four different parameter values, and taking 2.1 as the parameter value, the third sub-dataset after splitting is [1600, 2.1, 3.5], and the fourth sub-dataset is [1700, 2.5, 4.5], [1650, 2.8, 4.0], and [1750, 3.0, 5.0]. Then, according to the third and fourth sub-datasets, the total SSE of the third and fourth sub-datasets is calculated. By analogy, assuming that in this example, the total SSE of the third and fourth sub-datasets calculated after splitting with 2.5 as the parameter value is greater than the total SSE of the third and fourth sub-datasets generated after splitting with 2.1, 2.8, and 3.0 as the parameter values, then the density parameter of 2.5 is selected as the target parameter value of the density logging parameter.

[0109] It can be understood that by selecting the parameter value corresponding to the minimum value of the total SSE of the third sub-dataset and the fourth sub-dataset generated after splitting as the target parameter value, the accuracy of the shear wave velocity prediction value generated within the target parameter value area is guaranteed.

[0110] Step d, splitting the second data set of the known well into two new second data sets according to the target parameter value;

[0111] The new second data set is two similar data sets formed by splitting the second data set of the known well according to the target parameter value.

[0112] Continuing with the previous example, if the second data set of the known wells in Table 2 is split according to the density parameter of 2.5 as the target parameter value, two new second data sets are obtained as shown in Tables 3 and 4.

[0113] Table 3 New second data set 1 for a known well

[0114] depth density Natural Gamma Shear wave velocity 1600 2.1 30 3.5 1700 2.5 28 4.5

[0115] Table 4 New second data set 2 for a known well

[0116] depth density Natural Gamma Shear wave velocity 1650 2.8 32 4.0 1750 3.0 29 5.0

[0117] Step e, determining the logging parameters to be processed as processed logging parameters;

[0118] The processed logging parameter refers to the logging parameter that has been processed by splitting calculation and the corresponding target parameter value has been selected.

[0119] It can be understood that the logging parameters to be processed are determined as processed logging parameters to further process other subsequent target logging parameters. This method enables the decision tree model to process the logging parameters more quickly, avoids repeated processing of logging parameters, and improves the efficiency of model operation.

[0120] Repeat steps a to e until all target logging parameters have been processed.

[0121] Among them, by repeatedly executing steps a to e, the target parameter values ​​of all target logging parameters are determined using recursive logic, that is, the leaf nodes of the decision tree model constructed for each known well are determined, and finally the second data set is divided into multiple new second data sets, so that the corresponding decision tree model of the known well can be further constructed.

[0122] Optionally, assume that the given training data set is D = {(X 1 ,Y 1 ),(X 2 ,Y 2 ),…,(X N ,Y N )}, i.e. the first data set. is the parameter value of all logging parameters of each known well at each preset depth, G is the number of logging parameters. i=1,2,…,N, N is the sample size of the known well, Y i is the corresponding shear wave velocity value. When performing the regression task, the first data set space before the decision tree is split is defined as Q. When the decision tree is growing, the classification and regression tree algorithm is used to continuously divide the space Q into M split data set spaces Q 1 ,Q 2 ,…,Q m The predicted shear wave velocity corresponding to the split data set space is defined as C m , then the final decision tree model corresponding to the formula is:

[0123]

[0124] Where f(X) is the decision tree model formed by the known wells; M is the number of data sets after the known wells are split according to the target parameter value, C m is the shear wave velocity output from the data set after splitting according to the target parameter value.

[0125] Furthermore, in the process of step c above, the purpose of calculating the target parameter value of each target logging parameter is to minimize the sum of square errors of the shear wave velocity prediction values ​​corresponding to each data set space obtained by splitting according to the target logging parameter, thereby ensuring the prediction accuracy of the prediction model. The specific sum of square error formula is:

[0126] L=∑(Y i -f(X i )) 2 ,i=1,2,...,N

[0127] Where L is the square error of the predicted shear wave velocity values ​​corresponding to each data set space obtained by splitting according to the target logging parameters. When the square error L is the smallest, Q is taken. m The corresponding shear wave velocity Y in the split data set space i The average value of is taken as the predicted value of the shear wave velocity of the spatial output of the split data set.

[0128] It should be understood that by constructing the corresponding decision tree model for the known wells, a complete random forest model of the known wells is generated. In the prior art, the number of decision tree models established in the random forest model is usually set according to the network parameters, but in the embodiment of the present application, the number of decision tree models is consistent with the number of known wells. In this way, it can be ensured that each known well corresponds to a decision tree model. The uncertainty and error caused by random extraction in the random forest are avoided, and the stability and accuracy of the prediction model are ensured.

[0129] It can be understood that since the second data set includes different target logging parameters of known wells, and each target logging parameter contains different parameter values, the target parameter value of each target logging parameter is determined one by one using recursive logic, which effectively improves the model operation efficiency. At the same time, the target parameter value selection process also ensures the prediction accuracy of the decision tree model of each known well, thereby further ensuring the prediction accuracy of the random forest model composed of multiple known well decision trees.

[0130] Figure 6 Schematic diagram of the process of the shear wave velocity prediction method provided in the embodiment of the present application Figure 2 This embodiment is based on the above embodiment and describes in detail the process of determining the target logging parameters. Figure 6 As shown, the method includes:

[0131] S601, for each well logging parameter, extracting a third data set corresponding to the well logging parameter from a first data set of a known well;

[0132] The third data set includes the shear wave velocity of the known well at each preset depth and the parameter values ​​of the logging parameters;

[0133] Exemplarily, Table 5 is a third data set corresponding to natural gamma logging parameters extracted from the first data set of known wells.

[0134] Table 5 The third data set corresponding to the natural gamma logging parameters of a known well

[0135] depth Natural Gamma Shear wave velocity 1600 30 3.5 1650 32 4.0 1700 28 4.5 1750 29 5.0

[0136] It can be understood that extracting the third data set corresponding to the logging parameters from the first data set of the known well provides a data basis for the subsequent calculation of the cumulative reduction value of the MSE of the logging parameters.

[0137] S602, for each parameter value of the logging parameter, splitting a third data set corresponding to the logging parameter according to the parameter value to generate a first sub-data set and a second sub-data set;

[0138] The first sub-dataset and the second sub-dataset refer to two similar data sets formed after the third data set is split according to each parameter value target of the logging parameter.

[0139] It is understandable that the specific process of splitting the third data set corresponding to the logging parameters according to the parameter values ​​to generate the first sub-data set and the second sub-data set in S202 is similar to the splitting process in the above step c, which is not elaborated in this application.

[0140] S603, calculating the mean squared error (MSE) reduction value of the logging parameters before and after splitting according to the parameter values ​​according to the first sub-data set, the second sub-data set and the third data set corresponding to the logging parameters;

[0141] Among them, MSE is one of the commonly used indicators to measure the prediction performance of the model. The accuracy of the model is mainly evaluated by calculating the average value of the square of the error. The MSE reduction value refers to the difference between the MSE of the logging parameter before splitting according to the parameter value and the MSE of the logging parameter after splitting according to the parameter value.

[0142] It can be understood that by using the first sub-dataset to calculate the MSE of the first sub-dataset, using the second sub-dataset to calculate the MSE of the second sub-dataset, and using the third data set corresponding to the logging parameters to calculate the MSE before splitting according to the parameter value, and then combining these three MSE values, the MSE reduction value of the logging parameters before and after splitting according to the parameter value can be calculated.

[0143] In one achievable manner, the specific manner of calculating the mean square error (MSE) reduction value of the logging parameters before and after the splitting according to the parameter value can be achieved by the following method:

[0144] First, according to the third data set corresponding to the logging parameter, the first MSE of the logging parameter before splitting according to the parameter value is determined. Then, according to the first sub-data set and the second sub-data set, the second MSE of the logging parameter after splitting according to the parameter value is determined. Finally, the difference between the first MSE and the second MSE is determined as the MSE reduction value of the logging parameter before and after splitting according to the parameter value.

[0145] Optionally, the specific calculation formula of MSE is:

[0146]

[0147] Where k is the number of well shear wave velocities in different data sets; r i is the shear wave velocity of the known wells in the data set; r^ is the average shear wave velocity of all known wells in the data set.

[0148] Optionally, the specific calculation formula of the MSE reduction value is:

[0149]

[0150] Where ΔI(j,s) is the MSE reduction value generated by splitting the logging parameter j; I(s) is the first MSE; k s is the number of shear wave velocities in the third data set, k sleft is the number of shear wave velocities in the first sub-dataset; k sright is the number of shear wave velocities in the second sub-dataset; I(sleft) is the MSE calculated and generated based on the second sub-dataset; I(sright) is the MSE calculated and generated based on the second sub-dataset.

[0151] For example, assuming that the natural gamma in Table 5 is 28 as the parameter value for splitting, the first sub-dataset is [1700, 28, 4.5], [1750, 29, 5.0], and the second sub-dataset is [1650, 32, 4.0], [1600, 30, 3.5]. Then the r^ in the first MSE is the average value of all the shear wave data in the third data set, 4.25, k is 4, r i They are 3.5, 4.0, 4.5, and 5.0 respectively. The first MSE can be obtained by calculating through the MSE formula. Similarly, the MSE calculated based on the first sub-dataset, the MSE calculated based on the second sub-dataset, and the corresponding second MSE can be calculated. Finally, by calculating the difference between the first MSE and the second MSE, the MSE reduction value before and after the parameter value 28 split can be determined when natural gamma is used as the logging parameter.

[0152] It can be understood that by calculating the mean square error MSE corresponding to different data sets before and after the parameter value splitting of the logging parameters and combining them, the mean square error MSE reduction value of the logging parameters before and after the parameter value splitting can be determined, which can provide a basis for subsequently obtaining the complete mean square error MSE reduction value of the logging parameters and further quantify the degree of influence of the logging parameters on the shear wave velocity.

[0153] S604, calculating the MSE cumulative reduction value of the logging parameter according to the MSE reduction value corresponding to each parameter value of the logging parameter;

[0154] The MSE cumulative reduction value refers to a value obtained by accumulating the MSE reduction value corresponding to each parameter value of the logging parameter.

[0155] Optionally, the specific calculation formula for the MSE cumulative reduction value of the logging parameter is:

[0156]

[0157] Among them, I j is the cumulative decrease of MSE of logging parameter j.

[0158] It can be understood that by accumulating the MSE reduction value corresponding to each parameter value of the logging parameter, the MSE cumulative reduction value corresponding to the logging parameter can be calculated.

[0159] S605 . Determine a plurality of target logging parameters from all the logging parameters according to the MSE cumulative reduction value of each logging parameter.

[0160] It can be understood that the greater the MSE cumulative reduction value of the logging parameter, the greater the contribution of the logging parameter to reducing the error of the model predicted shear wave velocity, and thus it can be selected as the target logging parameter. Therefore, it is also necessary to sort the calculated MSE cumulative reduction value of each logging parameter to further determine multiple target logging parameters.

[0161] In one achievable method, a specific method for selecting multiple target logging parameters is as follows:

[0162] All logging parameters are sorted in descending order according to the cumulative decrease value of MSE, and the first preset number of logging parameters are determined as target logging parameters.

[0163] The preset number is a specific number that is pre-set or specified according to actual conditions before the model is built. The preset number is a positive integer greater than or equal to 1.

[0164] Optionally, in order to more conveniently select the logging parameters with a high degree of influence on the shear wave velocity, the MSE cumulative reduction values ​​corresponding to all logging parameters can also be normalized to obtain the importance score of each logging parameter, and the sum of the importance scores of all logging parameters is 1.

[0165] Optionally, the formula for normalizing the MSE cumulative reduction values ​​of different logging parameters is:

[0166]

[0167] Among them, NormalizedI j is the importance score of each logging parameter, ∑ G I G The sum of the cumulative reduction values ​​of the MSE of G logging parameters. By selecting a preset number of g (g = 1, 2, 3..., g < G) logging parameters ranked in the first order of importance, the parameter values ​​corresponding to these logging parameters are used as the input of the decision tree model corresponding to the known wells to complete the selection of the target logging parameters.

[0168] Figure 7(a) shows the importance score ranking of the logging parameters of the known well A1, Figure 7(b) shows the importance score ranking of the logging parameters of the known well A2, and Figure 7(c) shows the importance score ranking of the logging parameters of the known well A3. It can be seen that in the importance ranking of the logging parameters of the known well A1, Velocity P-wave (Vp)>Spontaneous Potential (SP)>Denity (DEN)>Compensated Neutron Log (CNL)>Shale Content (SH). The importance ranking of the logging parameters of the known well A2 is Vp>CNL>SP>SH>DEN. The importance ranking of the logging parameters of the known well A3 is Vp>CNL>SH>SP>DEN. These rankings represent the degree of influence of each logging parameter on the model prediction results, which decreases from high to low. Therefore, for known well A1, P-wave velocity and natural potential should be given priority when building the model, while for known wells A2 and A3, more attention should be paid to P-wave velocity and neutron logging parameters. By ranking the importance scores of logging parameters for known wells A1, A2, and A3, the influence of each parameter on the model prediction results is intuitively displayed. This provides a powerful parameter screening basis for S-wave velocity prediction and helps to improve the performance of the prediction model.

[0169] Accordingly, Figure 8The linear correlation heat map between each well logging parameter of the well to be predicted provided in the embodiment of the present application. The linear correlation heat map is a visualization tool for displaying the linear correlation between multiple variables. By using colors to represent the size and direction of the correlation coefficient, the relationship between the variables can be observed intuitively. The correlation coefficient ranges from -1 to 1. Figure 8 As shown in the figure, A4 is the well to be predicted. The result of the linear correlation of each logging parameter on the shear wave velocity in the well to be predicted A4 is: Vp>CNL>DEN>SP>SH. This shows that Vp has the strongest linear correlation with the shear wave velocity, while CNL has a strong linear correlation with the shear wave velocity, DEN has a general linear correlation with the shear wave velocity, and SP and SH have a weak linear correlation with the shear wave velocity. Therefore, for the well to be predicted A4, the decision tree model of the known well A1 will output a more accurate shear wave velocity prediction result.

[0170] Furthermore, suppose there are q known wells and one well to be tested in the work area. p . Set each training well W 1 ,W 2 ,...,W q As a separate data set, the total data set is now H = {D 1 ,D 2 ,...,D q}. For a known well W 1 ,W 2 ,...,W q The data set is modeled separately. The separate modeling process is to build a decision tree model f(x) for each known well and generate a decision tree model for the known well. When the well to be tested W is input into the decision tree model p After the corresponding target logging parameters are obtained, each known well decision tree model will obtain the well to be tested W p The corresponding prediction results

[0171] It can be understood that sorting the MSE cumulative reduction value of each logging parameter from high to low and selecting a certain number of logging parameters can reduce the error caused by low-impact logging parameters during the modeling process. Given that the distribution of logging data between known wells and wells to be predicted may be different, this method can better adapt to these differences and more accurately predict the shear wave velocity of the well to be predicted. Especially in complex geological conditions and when the logging data between wells are quite different, this method can greatly improve the model prediction accuracy.

[0172] The present application provides a shear wave velocity method, device and equipment. First, each parameter value of the logging parameter is split according to the first data set to obtain different similar sub-data sets. Then, the cumulative MSE reduction value corresponding to the logging parameter value is calculated according to the data set before the split and the different similar sub-data sets after the split; finally, the cumulative MSE reduction value corresponding to each logging parameter value is sorted, and a certain number of logging parameters are selected as target logging parameters. It can provide logging parameters with more significant effects for the decision tree model established for each subsequent known well, and the strategy of independently establishing a decision tree model for each known well can effectively capture the subtle differences between each known well, avoiding the modeling error caused by the difference in the distribution of logging parameters, and on the basis of improving the model prediction accuracy, it is more in line with the actual geological situation.

[0173] Fig. 9 Schematic diagram of the interface of the shear wave velocity prediction method provided in the embodiment of the present application; Fig. 9 As shown, in Fig. 9 In the spatial constraint process, the black dots represent the locations of multiple known wells, and the gray dots represent the locations of the wells to be predicted. First, the target logging parameters are screened for each known well, and the input logging parameters are sorted by the decision tree attribute importance score. The first two parameters with higher scores are selected as the input of the decision tree model of each known well, and a single known well model is established. In the prediction process, the logging parameters of the well to be predicted are input into the decision tree model of each known well to obtain their respective prediction results. Subsequently, the distance between each known well and the well to be predicted is calculated, and the inverse distance weighting strategy is applied to assign the inverse of the distance as a weight to the prediction result of each known well decision tree model. This method effectively utilizes the spatial information correlation between wells, reduces the impact of distant known wells on the prediction results, and improves the prediction accuracy. Finally, the weighted prediction results are summed to obtain the final shear wave velocity prediction result of the well to be predicted. Compared with the traditional machine learning shear wave velocity prediction algorithm, this application has higher accuracy and reliability. By selecting the target logging parameters and distance weighting method for each known well, it can more accurately reflect the underground geological conditions and establish a more accurate shear wave velocity prediction model.

[0174] Fig.10 A schematic diagram of the structure of a shear wave velocity prediction device provided in an embodiment of the present application; Fig.10 As shown, the shear wave velocity prediction device 100 comprises:

[0175] A determination module 1001 is used to determine, for each known well, a plurality of target logging parameters from all logging parameters according to a first data set of the known well, wherein the first data set includes the shear wave velocity of the known well at each preset depth and a parameter value of each logging parameter, and the influence of the target logging parameter on the shear wave velocity of the known well is greater than the influence of other logging parameters on the shear wave velocity of the known well;

[0176] A construction module 1002 is used to construct a decision tree model of the known well according to a second data set of the known well, where the second data set includes the shear wave velocity of the known well at each preset depth and the parameter value of each target logging parameter;

[0177] An input module 1003 is used to input a prediction data set of the prediction well into a decision tree model of a known well to obtain an initial shear wave velocity output by the decision tree model, wherein the prediction data set includes a parameter value of each target logging parameter of the prediction well at each preset depth;

[0178] The generating module 1004 is used to generate the target shear wave velocity of the well to be predicted according to the initial shear wave velocity output by the known well in each decision tree model.

[0179] In a possible implementation, the determination module 1001 is specifically configured to:

[0180] For each well logging parameter, extracting a third data set corresponding to the well logging parameter from the first data set of the known well, the third data set including the shear wave velocity of the known well at each preset depth and the parameter value of the well logging parameter;

[0181] For each parameter value of the logging parameter, splitting the third data set corresponding to the logging parameter according to the parameter value to generate a first sub-data set and a second sub-data set;

[0182] According to the first sub-data set, the second sub-data set and the third data set corresponding to the logging parameter, calculating the MSE reduction value of the logging parameter before and after splitting according to the parameter value;

[0183] Calculate the MSE cumulative reduction value of the logging parameter according to the MSE reduction value corresponding to each parameter value of the logging parameter;

[0184] According to the MSE cumulative reduction value of each logging parameter, multiple target logging parameters are determined from all the logging parameters.

[0185] In a possible implementation, the determination module 1001 is specifically configured to sort all logging parameters in descending order according to the cumulative reduction value of MSE, and determine the first preset number of logging parameters as target logging parameters.

[0186] In a possible implementation, the determination module 1001 is specifically configured to:

[0187] Determining, based on a third data set corresponding to the logging parameter, a first MSE of the logging parameter before splitting according to the parameter value;

[0188] Determine, based on the first sub-data set and the second sub-data set, a second MSE of the logging parameter after splitting according to the parameter value;

[0189] The difference between the first MSE and the second MSE is determined as the MSE reduction value of the logging parameter before and after splitting according to the parameter value.

[0190] In a possible implementation, the shear wave velocity prediction device further includes a processing module, which is specifically configured to:

[0191] Step a, determining the well logging parameters to be processed from the target well logging parameters corresponding to the known wells;

[0192] Step b, extracting a fourth data set corresponding to the well logging parameter to be processed from the second data set of the known well, the fourth data set including the shear wave velocity of the known well at each preset depth and the parameter value of the well logging parameter to be processed;

[0193] Step c, according to the fourth data set, determining a target parameter value from the parameter values ​​of the well logging parameters to be processed, and the total SSE of the third sub-dataset and the fourth sub-dataset generated after splitting the fourth data set according to the target parameter value is less than the total SSE of the third sub-dataset and the fourth sub-dataset generated after splitting the fourth data set according to other parameter values;

[0194] Step d, splitting the second data set of the known well into two new second data sets according to the target parameter value;

[0195] Step e, determining the logging parameters to be processed as processed logging parameters;

[0196] Repeat steps a to e until all target logging parameters have been processed.

[0197] In a possible implementation, the determination module 1001 is specifically configured to:

[0198] For each parameter value of the well logging parameter to be processed, splitting the fourth data set according to the parameter value to generate a third sub-data set and a fourth sub-data set;

[0199] Calculate the total SSE of the third sub-dataset and the fourth sub-dataset based on the third sub-dataset and the fourth sub-dataset;

[0200] The parameter value with the minimum total SSE is determined as the target parameter value.

[0201] In a possible implementation, the generating module 1004 is specifically configured to:

[0202] Obtaining the location information of each known well and the location information of the well to be predicted;

[0203] According to the initial shear wave velocity output by the known wells in each decision tree model, the location information of the known wells and the location information of the wells to be predicted, the target shear wave velocity of the wells to be predicted is generated.

[0204] In a possible implementation, the generating module is specifically used for:

[0205] By formula Generate target shear wave velocity of the well to be predicted;

[0206] Among them, y pred is the target shear wave velocity of the well to be predicted, is the initial shear wave velocity output by the decision tree model of the known well i, n is the number of known wells, b is the weighted power index, x is the horizontal coordinate of the well to be predicted, y is the vertical coordinate of the well to be predicted, and x i is the horizontal coordinate of the known well i, y i is the ordinate of the known well i. The abscissa and ordinate of the well to be predicted are determined according to the location information of the well to be predicted, and the abscissa and ordinate of the known well i are determined according to the location information of the known well i.

[0207] The shear wave velocity prediction device provided in the embodiment of the present application can be used to execute the shear wave velocity prediction method in any of the above embodiments. Its implementation principle and technical effects are similar and will not be repeated here.

[0208] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software calling through processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software calling through processing elements, and some modules can be implemented in the form of hardware. In addition, all or part of these modules can be integrated together, or they can be implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0209] Fig.11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Fig.11As shown, the electronic device 110 provided in this embodiment includes: at least one processor 1101 and a memory 1102. The device 110 also includes a communication component 1103. The processor 1101, the memory 1102 and the communication component 1103 are connected via a bus 1104.

[0210] In a specific implementation process, at least one processor 1101 executes the computer-executable instructions stored in the memory 1102, so that at least one processor 1101 executes the above-mentioned shear wave velocity prediction method.

[0211] The specific implementation process of the processor 1101 can be found in the above method embodiment, and its implementation principle and technical effect are similar, so this embodiment will not be repeated here.

[0212] The electronic device provided in the embodiment of the present application can be used to execute the shear wave velocity prediction method provided in any of the above method embodiments. Its implementation principle and technical effect are similar and will not be repeated here.

[0213] In the above Fig.11 In the illustrated embodiment, it should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the invention may be directly implemented as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor.

[0214] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (Non-volatile Memory, NVM), such as at least one disk memory.

[0215] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of the present application is not limited to only one bus or one type of bus.

[0216] Finally, it should be noted that those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not disclosed in the present application, are not limited to the precise structures described above and shown in the drawings, and may be modified and changed in various ways without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for predicting shear wave velocity, characterized in that: include: For each known well, a plurality of target logging parameters are determined from all logging parameters according to a first data set of the known well, wherein the first data set includes the shear wave velocity of the known well at each preset depth and the parameter value of each logging parameter, and the influence of the target logging parameter on the shear wave velocity of the known well is greater than the influence of other logging parameters on the shear wave velocity of the known well; constructing a decision tree model of the known well according to a second data set of the known well, wherein the second data set includes the shear wave velocity of the known well at each preset depth and the parameter value of each target logging parameter; Inputting a data set to be predicted of the well to be predicted into the decision tree model of the known well to obtain an initial shear wave velocity output by the decision tree model, wherein the data set to be predicted includes a parameter value of each target logging parameter of the well to be predicted at each preset depth; According to the initial shear wave velocity output by the known well in each decision tree model, the target shear wave velocity of the well to be predicted is generated.

2. The method according to claim 1, characterized in that For each known well, a plurality of target logging parameters are determined from all logging parameters according to the first data set of the known well, including: For each well logging parameter, extracting a third data set corresponding to the well logging parameter from the first data set of the known well, the third data set including the shear wave velocity of the known well at each preset depth and the parameter value of the well logging parameter; For each parameter value of the logging parameter, splitting a third data set corresponding to the logging parameter according to the parameter value to generate a first sub-data set and a second sub-data set; Calculate, according to the first sub-data set, the second sub-data set and a third data set corresponding to the logging parameter, a mean square error (MSE) reduction value of the logging parameter before and after splitting according to the parameter value; Calculating the MSE cumulative reduction value of the logging parameter according to the MSE reduction value corresponding to each parameter value of the logging parameter; According to the MSE cumulative reduction value of each logging parameter, a plurality of target logging parameters are determined from all the logging parameters.

3. The method according to claim 2, characterized in that Determining a plurality of target logging parameters from all the logging parameters according to the MSE cumulative reduction value of each logging parameter includes: All logging parameters are sorted in descending order according to the cumulative decrease value of MSE, and the first preset number of logging parameters are determined as the target logging parameters.

4. The method according to claim 2, characterized in that: The calculating, according to the first sub-data set, the second sub-data set and the third data set corresponding to the logging parameter, the MSE reduction value of the logging parameter before and after the splitting according to the parameter value comprises: determining, according to a third data set corresponding to the logging parameter, a first MSE of the logging parameter before splitting according to the parameter value; Determining, based on the first sub-dataset and the second sub-dataset, a second MSE of the logging parameter after splitting according to the parameter value; The difference between the first MSE and the second MSE is determined as the MSE reduction value of the logging parameter before and after the splitting according to the parameter value.

5. The method according to any one of claims 1 to 4, characterized in that: The step of constructing a decision tree model of the known well according to the second data set of the known well comprises: Step a, determining the well logging parameters to be processed from the target well logging parameters corresponding to the known wells; Step b, extracting a fourth data set corresponding to the well logging parameter to be processed from the second data set of the known well, wherein the fourth data set includes the shear wave velocity of the known well at each preset depth and the parameter value of the well logging parameter to be processed; Step c, according to the fourth data set, determining a target parameter value from the parameter values ​​of the to-be-processed logging parameters, wherein the total sum of square errors (SSE) of the third sub-dataset and the fourth sub-dataset generated after splitting the fourth data set according to the target parameter value is less than the total SSE of the third sub-dataset and the fourth sub-dataset generated after splitting the fourth data set according to other parameter values; Step d, splitting the second data set of the known well into two new second data sets according to the target parameter value; Step e, determining the to-be-processed logging parameters as processed logging parameters; Repeat step a to step e until all target logging parameters have been processed.

6. The method according to claim 5, characterized in that Determining a target parameter value from the parameter values ​​of the to-be-processed logging parameters according to the fourth data set includes: For each parameter value of the to-be-processed logging parameter, split the fourth data set according to the parameter value to generate a third sub-data set and a fourth sub-data set; Calculating the total SSE of the third sub-dataset and the fourth sub-dataset according to the third sub-dataset and the fourth sub-dataset; The parameter value with the minimum total SSE is determined as the target parameter value.

7. The method according to any one of claims 1 to 4, characterized in that: The step of generating the target shear wave velocity of the well to be predicted according to the initial shear wave velocity output by the known well in each decision tree model comprises: Obtaining the location information of each known well and the location information of the well to be predicted; According to the initial shear wave velocity output by the known wells in each decision tree model, the location information of the known wells and the location information of the wells to be predicted, the target shear wave velocity of the wells to be predicted is generated.

8. The method according to claim 7, characterized in that The method of generating a target shear wave velocity of the well to be predicted according to the initial shear wave velocity output by the known well in each decision tree model, the location information of the known well and the location information of the well to be predicted comprises: By formula generating a target shear wave velocity of the well to be predicted; Among them, y pred is the target shear wave velocity of the well to be predicted, is the initial shear wave velocity output by the decision tree model of the known well i, n is the number of known wells, b is the weighted power index, x is the horizontal coordinate of the well to be predicted, y is the vertical coordinate of the well to be predicted, and x i is the horizontal coordinate of the known well i, y i is the ordinate of the known well i. The abscissa and ordinate of the well to be predicted are determined according to the position information of the well to be predicted, and the abscissa and ordinate of the known well i are determined according to the position information of the known well i.

9. A shear wave velocity prediction device, characterized in that: include: a determination module, configured to determine, for each known well, a plurality of target logging parameters from all logging parameters according to a first data set of the known well, wherein the first data set includes the shear wave velocity of the known well at each preset depth and a parameter value of each logging parameter, wherein the target logging parameter has a greater influence on the shear wave velocity of the known well than other logging parameters; A construction module, configured to construct a decision tree model of the known well according to a second data set of the known well, wherein the second data set includes a shear wave velocity of the known well at each preset depth and a parameter value of each target logging parameter; An input module, used for inputting the to-be-predicted data set of the to-be-predicted well into the decision tree model of the known well, and obtaining the initial shear wave velocity output by the decision tree model; The generating module is used to generate the target shear wave velocity of the well to be predicted according to the initial shear wave velocity output by the known well in each decision tree model.

10. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the shear wave velocity prediction method according to any one of claims 1 to 8.

Citation Information

Patent Citations

  • Complex lithologic identification method based on decision tree data mining algorithm and complex lithologic identification system thereof

    CN107133670A

  • Method for inverting logging shear wave time difference by regression-fractal interpolation method

    CN110989034A

  • Transverse wave prediction method and system based on Gaussian process regression

    CN113466962A

  • Construction method and device of microseismic velocity model and electronic equipment

    CN114460635A

  • Transverse wave velocity prediction method, device and equipment of oil and gas well and storage medium

    CN116027433A

Cited By

  • Transverse wave time difference prediction method for deep carbonate rock fault control reservoir

    CN121878817A