Method, device and equipment for predicting shear wave velocity of oil and gas well and storage medium
By performing correlation analysis and model optimization on logging data and shear wave velocity data from oil and gas wells, the shear wave velocity prediction model with the lowest error was selected, which solved the problem of low prediction accuracy in the existing technology and achieved higher accuracy shear wave velocity prediction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA UNIV OF PETROLEUM (BEIJING)
- Filing Date
- 2022-10-26
- Publication Date
- 2026-06-02
AI Technical Summary
In existing technologies, shear wave velocity prediction methods have low prediction accuracy when there are significant differences between actual application conditions and statistical laws.
By performing correlation analysis on logging data and shear wave velocity data of identified oil and gas wells, target logging data with a correlation higher than the preset value are selected, multiple shear wave velocity prediction models are trained, hyperparameters are optimized, and the model with the lowest error is selected for prediction.
This improves the accuracy of shear wave velocity prediction, ensuring that the prediction results are closer to the actual values.
Smart Images
Figure CN116027433B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of oil and gas well development technology, and in particular to a method, apparatus, equipment and storage medium for predicting shear wave velocity in oil and gas wells. Background Technology
[0002] Shear wave velocity and P-wave velocity are crucial for evaluating the lithology, physical properties, and oil-bearing capacity of oil and gas reservoirs, as well as for the rational design of oil and gas well development programs. P-wave velocity can usually be obtained through conventional logging methods, while shear wave velocity data is often obtained only from two or three wells within a single oil and gas well area due to high testing costs. Therefore, shear wave velocity data for most oil and gas wells needs to be predicted.
[0003] Currently, among the existing technologies, the commonly used methods for predicting shear wave velocity include the empirical method, which derives the predicted shear wave velocity based on the statistical regularities of a large amount of conventional well logging data.
[0004] However, the inventors have found that the existing technology has at least the following technical problems: the empirical method is derived from the statistical laws of a large amount of shear wave velocity data in well areas. When the actual application conditions differ greatly from the statistical laws, the prediction accuracy is prone to be low. Summary of the Invention
[0005] This application provides a method, apparatus, equipment, and storage medium for predicting shear wave velocity in oil and gas wells, which can improve the prediction accuracy of shear wave velocity in oil and gas wells.
[0006] In a first aspect, this application provides a method for predicting the shear wave velocity of oil and gas wells, including:
[0007] Acquire logging data and shear wave velocity data from identified oil and gas wells;
[0008] Correlation analysis is performed on the logging data and shear wave velocity data of the identified oil and gas wells to obtain first target logging data with a correlation higher than a preset value. The first target logging data includes training data for model training, verification data for model hyperparameter optimization, and test data for obtaining model prediction error.
[0009] The training data is input into multiple different shear wave velocity prediction models for training, resulting in multiple different initial shear wave velocity prediction models. The validation data is input into each initial shear wave velocity prediction model for hyperparameter optimization, resulting in a secondary shear wave velocity prediction model.
[0010] The test data is input into multiple different secondary shear wave velocity prediction models, and the shear wave velocity prediction value output by each of the multiple different secondary shear wave velocity prediction models is obtained. The secondary shear wave velocity prediction model with the lowest error in the shear wave velocity prediction value among the multiple different secondary shear wave velocity prediction models is output as the target shear wave velocity prediction model.
[0011] The logging data of the oil and gas well to be predicted is input into the target shear wave velocity prediction model, and the shear wave velocity of the oil and gas well to be predicted is output.
[0012] In one possible implementation, the correlation analysis of the logging data and shear wave velocity data of the identified oil and gas well includes:
[0013] The logging data of the identified oil and gas wells are classified according to data type to obtain multiple sets of logging sequences;
[0014] The correlation between each logging sequence and the shear wave velocity of the identified oil and gas well is calculated, and the correlation between each logging sequence and the shear wave velocity of the identified oil and gas well is output.
[0015] In one possible implementation, the correlation calculation between each logging sequence and the shear wave velocity of the identified oil and gas wells includes:
[0016] use Perform correlation calculations;
[0017] Where r is the correlation between any set of logging sequences obtained from the classification and the shear wave velocity of the identified oil and gas wells, n is the number of sample points in the identified oil and gas wells, and x i Let be the value of the logging sequence corresponding to the i-th sample point out of n sample points. y is the average value of the well logging sequence corresponding to n sample points. i The shear wave velocity corresponding to the i-th sample point in the determined oil and gas well. Let be the mean of the shear wave velocities of all sample points in the determined oil and gas wells, where n is a natural number and i is a natural number greater than 0 and not greater than n.
[0018] In one possible implementation, before inputting the first target logging data into multiple different shear wave velocity prediction models for training, the method further includes:
[0019] The first target logging data is normalized to obtain normalized data of the first target logging data.
[0020] In one possible implementation, the secondary shear wave velocity prediction model with the lowest shear wave velocity prediction error among the plurality of different secondary shear wave velocity prediction models is used as the target shear wave velocity prediction model, including:
[0021] The test data is input into each of the secondary shear wave velocity prediction models, and the predicted shear wave velocity value is output.
[0022] Error calculation is performed on the predicted shear wave velocity value output by each of the secondary shear wave velocity prediction models and the shear wave velocity of the determined oil and gas well to obtain the prediction error value of each of the secondary shear wave velocity prediction models.
[0023] The secondary shear wave velocity prediction model with the lowest prediction error value is output as the target shear wave velocity prediction model.
[0024] In one possible implementation, the error calculation for the predicted shear wave velocity value output by each of the secondary shear wave velocity prediction models and the shear wave velocity of the determined oil and gas well includes:
[0025] use Perform error calculation;
[0026] use Calculate the mean absolute error;
[0027] Where RMSE is the root mean square error, MAE is the mean absolute error, m is the number of sample points in the test data, and y s Let be the shear wave velocity of the s-th sample point out of m sample points, representing a known oil and gas well. This is the predicted value of the shear wave velocity at the s-th sample point.
[0028] In one possible implementation, before inputting the logging data of the oil and gas well to be predicted into the target shear wave velocity prediction model, the method further includes:
[0029] If it is determined that the oil and gas well to be predicted is located in a different well area from the oil and gas well already identified, then a correlation analysis is performed on the logging data and shear wave velocity data of the identified oil and gas well located in the same well area as the oil and gas well to be predicted, to obtain a second target logging data with a correlation higher than a preset value.
[0030] If it is determined that the second target logging data and the first target logging data are of the same type, then the logging data of the oil and gas well to be predicted is input into the target shear wave velocity prediction model.
[0031] Secondly, this application provides a shear wave velocity prediction device for oil and gas wells, comprising:
[0032] The acquisition module is used to acquire logging data and shear wave velocity data of identified oil and gas wells;
[0033] The data analysis module is used to perform correlation analysis on the logging data and shear wave velocity data of the identified oil and gas wells to obtain first target logging data with a correlation higher than a preset value. The first target logging data includes training data for model training, verification data for model hyperparameter optimization, and test data for obtaining model prediction errors.
[0034] The model training module is used to input the training data into multiple different shear wave velocity prediction models for training, thereby obtaining multiple different initial shear wave velocity prediction models. The validation data is then input into each initial shear wave velocity prediction model for hyperparameter optimization, thereby obtaining a secondary shear wave velocity prediction model.
[0035] The model training module is also used to input the test data into multiple different secondary shear wave velocity prediction models, obtain the shear wave velocity prediction value output by each secondary shear wave velocity prediction model, and output the shear wave velocity prediction model with the lowest shear wave velocity prediction value error among the multiple secondary shear wave velocity prediction models as the target shear wave velocity prediction model.
[0036] The prediction module is used to input the logging data of the oil and gas well to be predicted into the target shear wave velocity prediction model and output the shear wave velocity of the oil and gas well to be predicted.
[0037] Thirdly, this application provides a shear wave velocity prediction device for oil and gas wells, comprising: at least one processor and a memory;
[0038] The memory is used to store computer-executed instructions;
[0039] The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in the first aspect.
[0040] Fourthly, this application provides a computer-readable storage medium storing an executable instruction of a computer, which, when executed by a processor, implements the method described in the first aspect.
[0041] This application provides a method, apparatus, device, and storage medium for predicting shear wave velocity in oil and gas wells. The method involves performing correlation analysis on logging data and shear wave velocity data of a known oil and gas well to obtain first target logging data with a correlation higher than a preset value. Multiple different initial shear wave velocity prediction models are then trained using training data. Validation data is used to optimize the hyperparameters of these initial models, resulting in multiple optimized secondary shear wave velocity prediction models. Test data is then used to analyze the error of the predicted shear wave velocity values from these secondary models, and the secondary model with the lowest prediction error is selected as the target shear wave velocity prediction model. Finally, the target shear wave velocity prediction model is used to predict the shear wave velocity of the oil and gas well to be predicted, thus improving the accuracy of the predicted shear wave velocity. Attached Figure Description
[0042] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0043] Figure 1 A schematic flowchart illustrating a method for predicting shear wave velocity in oil and gas wells, provided as an embodiment of this application;
[0044] Figure 2 This is a schematic diagram of conventional logging data and shear wave velocity data for oil and gas wells with a depth of 3808-5200m as defined in this application embodiment.
[0045] Figure 3 This is a schematic diagram of the shear wave velocity prediction structure provided in an embodiment of this application;
[0046] Figure 4 A schematic diagram of the structure of a shear wave velocity prediction device for an oil and gas well provided in an embodiment of this application;
[0047] Figure 5 This is a schematic diagram of the hardware structure of a shear wave velocity prediction device for oil and gas wells provided in an embodiment of this application. Detailed Implementation
[0048] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0049] Currently, shear wave velocity and p-wave velocity are key geophysical parameters for pre-stack inversion, reservoir characterization, and fluid distribution identification, crucial for evaluating reservoir lithology, physical properties, and oil-bearing potential, and for rationally designing development plans. Existing technologies typically employ empirical methods for shear wave velocity prediction, such as the Carroll algorithm, Han algorithm, and Poisson's ratio algorithm. These methods rely on extensive shear wave velocity data from explored oil and gas wells to summarize patterns and derive the shear wave velocity of the well to be predicted. The inventors have discovered that in existing technologies, because each well area generally contains hundreds of oil and gas wells, and because testing the shear wave velocity of oil and gas wells using dipole acoustic waves is costly, only two to three wells are selected per well area for shear wave velocity data testing. Therefore, using shear wave velocity data collected from multiple well areas for empirical prediction of the shear wave velocity of other wells to be predicted can lead to low prediction accuracy when the application conditions differ significantly from actual conditions.
[0050] To solve the above-mentioned technical problems, the embodiments of this application provide the following technical concept for solving the problems: First, a model is constructed and trained based on conventional logging data and shear wave velocity data. Then, conventional logging data of the oil and gas well to be predicted is input into the trained model. The trained model is optimized and tested and evaluated. The final prediction model is used to predict the shear wave velocity of the oil and gas well to be predicted, thereby improving the prediction accuracy.
[0051] Figure 1 This is a flowchart illustrating a method for predicting the shear wave velocity of an oil and gas well, provided in an embodiment of this application. The execution entity of this embodiment can be a device with data calculation and processing functions, such as various types of terminal devices or servers, etc. No particular limitation is made to this embodiment.
[0052] like Figure 1 As shown, the method for predicting the shear wave velocity of this oil and gas well includes:
[0053] S101: Obtain logging data and shear wave velocity data for identified oil and gas wells.
[0054] In this embodiment, the identified oil and gas wells can be those that have been exploited and whose geophysical data has been obtained. These wells can also be identified using shear wave velocity logging technology. For example, the logging data for oil and gas wells may include GR (Gamma Ray) data, Vp (compressional wave velocity) data, RT (resistivity) data, and Poro (neutron porosity) data, while the shear wave velocity data is the predicted data for the identified oil and gas wells. Please refer to [reference needed]. Figure 2 , Figure 2 This is a schematic diagram of logging data and shear wave velocity data for wells with confirmed depths of 3808-5200m.
[0055] S102: Perform correlation analysis on the logging data and shear wave velocity data of the identified oil and gas wells to obtain the first target logging data with a correlation higher than the preset value. The first target logging data includes training data for model training, verification data for model hyperparameter optimization, and test data for obtaining model prediction error.
[0056] In this embodiment, correlation analysis involves analyzing two or more correlated variable elements to measure the degree of correlation between the two variables. The correlation analysis yields a correlation value; the closer this value is to 1 or -1, the stronger the correlation between the two variables. Models trained with highly correlated variables exhibit higher accuracy. Therefore, correlation analysis is performed on logging data and shear wave velocity data from identified oil and gas wells. This step can employ Pearson correlation coefficient, Kendall correlation coefficient, or Spearman correlation coefficient algorithms.
[0057] In this embodiment, model training is a process based on machine learning that uses well logging data from the target data as input data and shear wave velocity as output data to iterate repeatedly to obtain a model that can predict shear wave velocity.
[0058] In this embodiment, hyperparameter optimization is the process of optimizing the hyperparameters of the shear wave velocity prediction model. The hyperparameters are parameters that are manually set by the model in each shear wave prediction model before the machine learning training process begins. For example, the hyperparameter of the random forest model is the number of decision trees, the hyperparameters of the support vector machine model include the penalty coefficient and the coefficient of the kernel function, and the hyperparameters of the recurrent neural network model, long short-term memory model, gated sequence unit model and stacked gated recurrent unit model all include the learning rate, the number of neurons, the number of layers and the sliding window size.
[0059] In this embodiment, the purpose of dividing the first target data is to improve the prediction accuracy of the final target shear wave velocity prediction model. For example, the ratio of training data, validation data, and test data in the first target logging data can be 75%:15%:15%.
[0060] S103: Input the training data into multiple different shear wave velocity prediction models for training to obtain multiple different initial shear wave velocity prediction models. Input the validation data into each initial shear wave velocity prediction model for hyperparameter optimization to obtain a secondary shear wave velocity prediction model.
[0061] In this embodiment, the multiple different shear wave velocity prediction models are various types of models used to predict shear wave velocity. These models can be random forest models, support vector machine models, recurrent neural network models, long short-term memory models, gated recurrent unit models, and stacked gated recurrent unit models.
[0062] In this embodiment, the initial shear wave velocity prediction models are models obtained after training multiple different shear wave velocity prediction models. The secondary shear wave velocity prediction model is a model after hyperparameter optimization.
[0063] S104: Input the test data into multiple different secondary shear wave velocity prediction models, obtain the shear wave velocity prediction value output by each secondary shear wave velocity prediction model, and output the secondary shear wave velocity prediction model with the lowest shear wave velocity prediction value error among multiple different secondary shear wave velocity prediction models as the target shear wave velocity prediction model.
[0064] In this embodiment, after the test data is input into each secondary shear wave velocity prediction model, a corresponding predicted shear wave velocity value is obtained. The prediction error is the value obtained by calculating the error between the predicted shear wave velocity value and the shear wave velocity data value of the determined oil and gas well. The error calculation method can be either the variance algorithm or the root mean square error algorithm. The lower the prediction error, the higher the prediction accuracy of the secondary shear wave velocity prediction model. Therefore, the shear wave velocity prediction model with the lowest predicted shear wave velocity value is output as the target shear wave velocity prediction model. The input of the target shear wave velocity prediction model is the logging data of the oil and gas well, and the output is the predicted shear wave velocity data.
[0065] S105: Input the logging data of the oil and gas well to be predicted into the target shear wave velocity prediction model, and output the shear wave velocity of the oil and gas well to be predicted.
[0066] In this embodiment, the oil and gas well to be predicted is an oil and gas well whose shear wave velocity has not been measured. The data type and acquisition method of the logging data for the oil and gas to be predicted are the same as those of the logging data for the wells whose oil and gas have been determined.
[0067] In summary, this application provides a method for predicting shear wave velocity in oil and gas wells. By performing correlation analysis on logging data and shear wave velocity data of a known oil and gas well, a first target logging data with a correlation higher than a preset value is obtained. Then, multiple different initial shear wave velocity prediction models are trained using training data. Next, hyperparameter optimization is performed on these multiple initial shear wave velocity prediction models using validation data to obtain multiple optimized secondary shear wave velocity prediction models. Finally, error analysis is performed on the predicted shear wave velocity values of these multiple secondary shear wave velocity prediction models using test data. The secondary shear wave velocity prediction model with the lowest prediction error is selected as the target shear wave velocity prediction model. Finally, the target shear wave velocity prediction model is used to predict the shear wave velocity of the oil and gas well to be predicted, resulting in higher accuracy in the predicted shear wave velocity.
[0068] Based on the foregoing embodiments, in step S102, when performing correlation analysis on the logging data and shear wave velocity of the identified oil and gas wells, in an optional implementation of this application, the specific steps of the correlation analysis include:
[0069] S102a: Classify the logging data of identified oil and gas wells according to data type to obtain multiple sets of logging sequences.
[0070] S102b: Perform correlation calculations between each logging sequence and the shear wave velocity of the identified oil and gas wells, and output the correlation between each logging sequence and the shear wave velocity of the identified oil and gas wells.
[0071] In this embodiment, the logging data consists of logging sequences collected from multiple sample points. The logging sequence of each sample point can have multiple types. For example, the logging data of all sample points can have data types such as GR, Vp, RT and Poro. Therefore, the GR data of all sample points can be used as a set of logging sequences, the Vp data of all sample points can be used as a set of logging sequences, and so on.
[0072] Based on the above embodiments, as an optional implementation of this application, in step S102b, correlation calculations are performed on each group of logging sequences and the shear wave velocities of the identified oil and gas wells. Specifically, the correlation calculations can be performed using... Perform correlation calculations.
[0073] Where r represents the correlation between any set of logging sequences obtained from the classification and the shear wave velocity of the identified oil and gas wells, n represents the number of sample points in the identified oil and gas wells, and x represents the correlation between r and r. i Let be the value of the logging sequence corresponding to the i-th sample point out of n sample points. y is the average value of the well logging sequence corresponding to n sample points. i Given the determined shear wave velocity at the i-th sample point in the oil and gas well, Let be the mean of the shear wave velocities at all sample points in the determined oil and gas wells, where n is a natural number and i is a natural number greater than 0 and not greater than n.
[0074] For example, please continue to refer to Figure 2 Assume that the number of sample points in the oil and gas wells is 10416, i.e., n = 10416, and i belongs to any one of the 10416 sample points. For example... Figure 2 As shown, the four types of logging data for oil and gas wells have been identified as GR, Vp, Poro, and RT. Correlation analysis is then performed on these four types of data and the shear wave velocity data. The correlation coefficients for each type of data are shown in Table 1 below.
[0075]
[0076] Table 1
[0077] As shown in Table 1, the correlation between P-wave velocity Vp and S-wave velocity is 0.738, and the correlation between neutron porosity Poro and S-wave velocity is -0.477. The correlations between P-wave velocity Vp and S-wave velocity, as well as between neutron porosity Poro and S-wave velocity, are relatively strong. Therefore, the P-wave velocity data and neutron porosity data from the logging data of the identified oil and gas wells are used as the first target logging data.
[0078] Based on the above embodiments, before inputting the first target logging data into multiple different shear wave velocity prediction models for training in step S103, as an alternative embodiment of this application, the shear wave velocity prediction method for oil and gas wells further includes the following steps: normalizing the first target logging data to obtain normalized data of the first target logging data.
[0079] In this embodiment, the normalized data refers to data within the range of 0 to 1 obtained after normalization processing. The formula used for normalization processing is:
[0080]
[0081] Where x is any data point in the first target logging data. For any data point in the first target logging data, x is the normalized data. min x is the smallest value among the data of the same type as any data in the first target logging data. max The data with the largest value among the data of the same type as any data in the first target logging data.
[0082] For example, if any P-wave velocity Vp in the first target logging data is x = 15 km / s, and the smallest value among all P-wave velocities Vp of the same type as x = 15 km / s is 13 km / s, and the largest value among all P-wave velocities Vp of the same type as x = 15 km / s is 17 km / s, then after normalization, the normalized data for x is: meters per second.
[0083] In summary, the shear wave velocity prediction method for oil and gas wells provided in this application further normalizes the first target logging data before inputting it into the shear wave velocity prediction model for training. This avoids the training failing to converge due to singular sample data when it appears in the first target logging data, thus increasing the training time.
[0084] Based on the above embodiments, as a feasible implementation method of this embodiment, when performing hyperparameter optimization in step S102, the hyperparameter optimization algorithm used can be a Bayesian optimization algorithm or a sparrow optimization algorithm. Taking the sparrow optimization algorithm as an example, a determined sample point is considered as one sparrow, and an oil and gas well has multiple sample points, so an oil and gas well is considered as one sparrow population. During each iteration of training the shear wave velocity prediction model, the sparrow optimization algorithm performs hyperparameter optimization to obtain some better hyperparameters. Hyperparameter optimization stops when the number of training iterations reaches the target value or the fitness value of each calculated sample point tends to stabilize. The last obtained hyperparameters are then used as the hyperparameters of the shear wave velocity prediction model, and the optimized target shear wave velocity prediction model is determined based on the last obtained hyperparameters.
[0085] Based on the above embodiments, the hyperparameters of the multiple initial shear wave velocity prediction models mentioned in step S103, such as the random forest model, support vector machine model, recurrent neural network model, long short-term memory model, and gated recurrent unit model, are optimized. Assuming that the number of sample points for a determined oil and gas well is 10416, then by calculating 15%, the resulting test dataset contains 1562 sample points. These 1562 sample points are input into multiple target shear wave prediction models for hyperparameter optimization, resulting in the optimized hyperparameter table of the multiple target shear wave prediction models, as shown in Table 2.
[0086]
[0087] Table 2
[0088] As shown in Table 2, the prediction results of each shear wave velocity prediction model after hyperparameter optimization are closer to the predicted results of the shear wave prediction model without hyperparameter optimization than the predicted results of the shear wave prediction model of the determined oil and gas well.
[0089] In an optional embodiment of this application, step S103 outputs the shear wave velocity prediction model with the lowest shear wave velocity prediction error among multiple secondary shear wave velocity prediction models as the target shear wave velocity prediction model, specifically including the following steps:
[0090] Step a: Input the test data into each secondary shear wave velocity prediction model and output the predicted shear wave velocity value.
[0091] Step b: Calculate the error between the predicted shear wave velocity output by each secondary shear wave velocity prediction model and the determined shear wave velocity of the oil and gas well to obtain the prediction error value of each secondary shear wave velocity prediction model.
[0092] Step c: Output the secondary shear wave velocity prediction model with the lowest prediction error as the target shear wave velocity prediction model.
[0093] In this embodiment, the predicted shear wave velocity value output by each shear wave velocity prediction model is only used to evaluate the prediction error calculation of that model. Specifically, in step b, when calculating the error between the predicted shear wave velocity value output by each secondary shear wave velocity prediction model and the shear wave velocity of the determined oil and gas well, a variance calculation formula or a standard deviation calculation formula is used.
[0094] As a feasible implementation method of the above embodiments, the prediction error calculation mentioned in the above embodiments can be adopted. Perform root mean square error calculation and use Calculate the mean absolute error. Here, RMSE is the root mean square error, MAE is the mean absolute error, m is the number of sample points in the test data, and y... s Let be the shear wave velocity of the s-th sample point out of m sample points, representing a known oil and gas well. This is the predicted value of the shear wave velocity at the s-th sample point.
[0095] For example, assuming an oil and gas well has 10,416 sample points, then calculating at 15%, the resulting test data sample points are 1,562, so m = 1,562. Continue to refer to... Figure 2 ,Will Figure 2 The 1562 sample points in the test dataset shown are input into each secondary shear wave velocity prediction model. Each secondary shear wave velocity prediction model will then obtain a predicted shear wave velocity. Originally, each sample point also had a corresponding transverse wave velocity y. s Then, by using the RMSE formula and MAE to calculate the error, the average index of the prediction results of each target shear wave velocity prediction model can be obtained, as shown in Table 3:
[0096] Evaluation indicators Root mean square error Mean Absolute Error Random Forest 0.25 0.22 Support Vector Machine 0.33 0.29 Recurrent Neural Networks 0.23 0.19 Long Short-Term Memory 0.20 0.13 Gated Loop Unit 0.10 0.05 Stacked gated loop unit 0.19 0.10
[0097] Table 3
[0098] As shown in Table 2, the root mean square error and mean absolute error of the gated cyclic unit model are the smallest, at 0.1 and 0.05 respectively, indicating that the prediction error of the gated cyclic unit model is the lowest. In this example, the gated cyclic unit model can be used as the target shear wave velocity prediction model output.
[0099] In summary, the shear wave velocity prediction method for oil and gas wells provided in this application embodiment also uses error calculation to enable the shear wave velocity prediction values of all sample points output by each secondary shear wave velocity prediction model to have a basis for judging whether the prediction error is high or low, which facilitates the selection of a secondary shear wave velocity prediction model with higher prediction accuracy as the target shear wave velocity prediction model.
[0100] In an optional embodiment of this application, the difference from the above embodiment is that, before step S104, the shear wave velocity prediction method for the oil and gas well further includes the following steps:
[0101] Step A: If it is determined that the oil and gas well to be predicted is located in a different well area from the identified oil and gas well, then a correlation analysis is performed on the logging data and shear wave velocity data of the identified oil and gas well located in the same well area as the oil and gas well to be predicted, to obtain second target logging data with a correlation higher than the preset value.
[0102] Step B: If it is determined that the second target logging data and the first target logging data are of the same type, then input the logging data of the oil and gas well to be predicted into the target shear wave velocity prediction model.
[0103] In this embodiment, if the oil and gas well to be predicted and the identified oil and gas well are located in different well areas, the logging data types strongly correlated with shear wave velocity may be different or the same. In this case, if the logging data of the oil and gas well to be predicted is directly input into the target velocity prediction model, the prediction accuracy of the obtained shear wave velocity will be reduced. In this embodiment, the process and method of correlation analysis are the same as the correlation analysis process in S102 of the above method embodiment, and will not be repeated here.
[0104] The above is part of the implementation method for predicting shear wave velocity when the oil and gas well to be predicted and the identified oil and gas well are in the same well area. The following describes in detail the process of predicting shear wave velocity when the oil and gas well to be predicted and the identified oil and gas well are located in different well areas, in conjunction with the implementation method.
[0105] In an optional embodiment of this application, if the oil and gas well to be predicted and the identified oil and gas well are located in different well areas, and the first target logging data and the second target logging data are of different types, then the logging data and shear wave velocity data of the identified oil and gas well located in the same well area as the oil and gas well to be predicted are used to train the model to obtain a target shear wave velocity prediction model suitable for the oil and gas well to be predicted, and then the shear wave velocity of the oil and gas well to be predicted is predicted. For the specific model training process and prediction process, please refer to steps S103 and S104 in the above method embodiment.
[0106] In summary, the shear wave velocity prediction method for oil and gas wells provided in this application first obtains second target logging data with a correlation higher than a preset value through correlation analysis before predicting the shear wave velocity of the oil and gas well to be predicted. After confirming that the second target logging data and the first target logging data are the same, the logging data of the oil and gas well to be predicted is then input into the target shear wave velocity prediction model. This avoids the influence of environmental factors on the prediction accuracy and makes the prediction accuracy of the obtained shear wave velocity higher.
[0107] Based on the above embodiments, this paper demonstrates the effectiveness of the shear wave velocity prediction method for oil and gas wells provided in the above embodiments by taking the gated circulation unit obtained in step S103 as the target shear wave velocity prediction model. Please refer to... Figure 3 , Figure 3 This is a schematic diagram of the shear wave velocity prediction structure provided in an embodiment of this application. Figure 3 As shown, the horizontal axis represents the depth of the oil and gas well to be predicted, and the vertical axis represents the shear wave velocity value corresponding to the depth. Clearly, compared to the shear wave velocity predicted by empirical methods, the shear wave velocity curve generated in this embodiment is closer to the actual shear wave velocity.
[0108] Please refer to Figure 4 , Figure 4 This is a schematic diagram of the structure of a shear wave velocity prediction device for oil and gas wells provided in an embodiment of this application. The device includes: an acquisition module 41, a data analysis module 42, a model training module 43, and a prediction module 44.
[0109] The acquisition module 41 is used to acquire logging data and shear wave velocity data of the identified oil and gas wells.
[0110] The data analysis module 42 is used to perform correlation analysis on the logging data and shear wave velocity data of the identified oil and gas wells to obtain first target logging data with a correlation higher than the preset value. The first target logging data includes training data for model training, verification data for model hyperparameter optimization, and test data for obtaining model prediction error.
[0111] The model training module 43 is used to input training data into multiple different shear wave velocity prediction models for training, thereby obtaining multiple different initial shear wave velocity prediction models. The validation data is then input into each initial shear wave velocity prediction model for hyperparameter optimization, thereby obtaining a secondary shear wave velocity prediction model.
[0112] The model training module 43 is also used to input test data into multiple different secondary shear wave velocity prediction models, obtain the shear wave velocity prediction value output by each secondary shear wave velocity prediction model, and output the shear wave velocity prediction model with the lowest shear wave velocity prediction value error among multiple secondary shear wave velocity prediction models as the target shear wave velocity prediction model.
[0113] The prediction module 44 is used to input the logging data of the oil and gas well to be predicted into the target shear wave velocity prediction model and output the shear wave velocity of the oil and gas well to be predicted.
[0114] The shear wave velocity prediction device for oil and gas wells provided in this embodiment can be used to execute the technical solution of the above method embodiment. Its implementation principle and the technical effect produced by the method embodiment are the same, and will not be described again in this embodiment.
[0115] In an optional embodiment of this application, the data analysis module 42 is specifically used to classify the logging data of the identified oil and gas wells according to the data type to obtain multiple sets of logging sequences, and to perform correlation calculations between each set of logging sequences and the shear wave velocity of the identified oil and gas wells, and output the correlation between each set of logging sequences and the shear wave velocity of the identified oil and gas wells.
[0116] In an optional embodiment of this application, when the data analysis module 42 performs correlation calculations on the shear wave velocities of each logging sequence and the identified oil and gas wells, it adopts... Perform correlation calculations.
[0117] Where r represents the correlation between any set of logging sequences obtained from the classification and the shear wave velocity of the identified oil and gas wells, n represents the number of sample points in the identified oil and gas wells, and x represents the correlation between r and r. i Let be the value of the logging sequence corresponding to the i-th sample point out of n sample points. Let be the average value of the logging sequence corresponding to n sample points, and yi be the shear wave velocity corresponding to the i-th sample point in the identified oil and gas well. Let be the mean of the shear wave velocities at all sample points in the determined oil and gas wells, where n is a natural number and i is a natural number greater than 0 and not greater than n.
[0118] In an optional embodiment of this application, the data analysis module 42 is further specifically used to normalize the first target logging data to obtain normalized data of the first target logging data.
[0119] In an optional embodiment of this application, the model training module 43 is further configured to input test data into each secondary shear wave velocity prediction model and output predicted shear wave velocity values. Specifically, the model training module 43 is further configured to calculate the error between the predicted shear wave velocity values output by each secondary shear wave velocity prediction model and the determined shear wave velocity of the oil and gas well, thereby obtaining the prediction error value of each secondary shear wave velocity prediction model. Specifically, the model training module 43 is further configured to output the secondary shear wave velocity prediction model with the lowest prediction error value as the target shear wave velocity prediction model.
[0120] In an optional embodiment of this application, the model training module 43 uses the following method when performing error calculation: Root mean square error (RMSE) calculation is performed. (Using...) Calculate the mean absolute error. Here, RMSE is the root mean square error, MAE is the mean absolute error, m is the number of sample points in the test data, and y... s Let be the shear wave velocity of the s-th sample point out of m sample points, representing a known oil and gas well. This is the predicted value of the shear wave velocity at the s-th sample point.
[0121] In an optional embodiment of this application, the data analysis module 42 is further configured to determine whether the oil and gas well to be predicted is located in a different well area from the identified oil and gas well. If the determination result is yes, then a correlation analysis is performed on the logging data and shear wave velocity data of the identified oil and gas well located in the same well area as the oil and gas well to be predicted, to obtain second target logging data with a correlation higher than a preset value. In this embodiment, the prediction module 44 is further configured to determine whether the second target logging data and the first target logging data are of the same type. If so, then the logging data of the oil and gas well to be predicted is input into the target shear wave velocity prediction model.
[0122] Please refer to Figure 5 , Figure 5 A schematic diagram of the hardware structure of a shear wave velocity prediction device for oil and gas wells provided in this application embodiment is shown below. Figure 5 As shown, the system includes at least one processor 501 and a memory 502.
[0123] The processor 501 is used to store computer execution instructions.
[0124] The memory 502 is used to execute computer execution instructions stored in the memory to implement the various steps involved in the above method embodiments. For details, please refer to the technical solutions described in the foregoing method embodiments.
[0125] Alternatively, the memory 502 can be either standalone or integrated with the processor 501.
[0126] When the memory 502 is set up independently, the controller also includes a bus 503 for connecting the memory 502 and the processor 501.
[0127] This invention also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the technical solution described in the above-described method embodiments.
[0128] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the technical solutions of the above-described method embodiments.
[0129] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between devices or modules, and may be electrical, mechanical, or other forms.
[0130] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to implement the solution of this embodiment according to actual needs.
[0131] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The unit composed of the above modules can be implemented in hardware or in the form of hardware plus software functional units.
[0132] The integrated modules described above, implemented as software functional modules, can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute partial steps of the methods of the various embodiments of this application.
[0133] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.
[0134] The memory may include high-speed RAM, and may also include non-volatile storage (NVM), such as at least one disk storage device, and may also be a USB flash drive, external hard drive, read-only memory, disk or optical disc, etc.
[0135] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.
[0136] The aforementioned storage medium can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium can be any available medium accessible to general-purpose or special-purpose computers.
[0137] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of the processor. Both the processor and the storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and storage medium can exist as discrete components in an electronic device or host device.
[0138] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.
[0139] This description is intended to illustrate the technical solutions of this application, and not to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.
Claims
1. A method for predicting shear wave velocity in oil and gas wells, characterized in that, include: Acquire logging data and shear wave velocity data from identified oil and gas wells; The logging data of the identified oil and gas wells are classified according to data type to obtain multiple sets of logging sequences; The correlation between each logging sequence and the shear wave velocity of the identified oil and gas wells is calculated, and the correlation between each logging sequence and the shear wave velocity of the identified oil and gas wells is output to obtain first target logging data with a correlation higher than a preset value. The first target logging data includes training data for model training, validation data for model hyperparameter optimization, and test data for obtaining model prediction errors. The correlation calculation between each logging sequence and the shear wave velocity of the identified oil and gas wells includes: use Perform correlation calculations; Where r represents the correlation between any set of logging sequences obtained from the classification and the shear wave velocity of the identified oil and gas wells, and n represents the number of sample points in the identified oil and gas wells. Let be the value of the logging sequence corresponding to the i-th sample point out of n sample points. This represents the average value of the logging sequences corresponding to n sample points. The shear wave velocity corresponding to the i-th sample point in the determined oil and gas well. Let be the mean of the shear wave velocities of all sample points in the determined oil and gas wells, where n is a natural number and i is a natural number greater than 0 and not greater than n; The training data is input into multiple different shear wave velocity prediction models for training, resulting in multiple different initial shear wave velocity prediction models. The validation data is input into each initial shear wave velocity prediction model for hyperparameter optimization, resulting in a secondary shear wave velocity prediction model. The test data is input into multiple different secondary shear wave velocity prediction models, and the shear wave velocity prediction value output by each secondary shear wave velocity prediction model is obtained. The shear wave velocity prediction model with the lowest error among the multiple secondary shear wave velocity prediction models is output as the target shear wave velocity prediction model. The logging data of the oil and gas well to be predicted is input into the target shear wave velocity prediction model, and the shear wave velocity of the oil and gas well to be predicted is output.
2. The method according to claim 1, characterized in that, Before inputting the first target logging data into multiple different shear wave velocity prediction models for training, the method further includes: normalizing the first target logging data to obtain normalized data of the first target logging data.
3. The method according to claim 1, characterized in that, The target shear wave velocity prediction model is the one with the lowest predicted shear wave velocity value error among the multiple secondary shear wave velocity prediction models, including: The test data is input into each of the secondary shear wave velocity prediction models, and the predicted shear wave velocity value is output. Error calculation is performed on the predicted shear wave velocity value output by each of the secondary shear wave velocity prediction models and the shear wave velocity of the determined oil and gas well to obtain the prediction error value of each of the secondary shear wave velocity prediction models. The secondary shear wave velocity prediction model with the lowest prediction error value is output as the target shear wave velocity prediction model.
4. The method according to claim 3, characterized in that, The error calculation for the predicted shear wave velocity output by each of the secondary shear wave velocity prediction models and the determined shear wave velocity of the oil and gas well includes: use Calculate the root mean square error; use Calculate the mean absolute error; Where RMSE is the root mean square error, MAE is the mean absolute error, and m is the number of sample points in the test data. Let be the shear wave velocity of the s-th sample point out of m sample points, representing a known oil and gas well. This is the predicted value of the shear wave velocity at the s-th sample point.
5. The method according to any one of claims 1 to 4, characterized in that, Before inputting the logging data of the oil and gas well to be predicted into the target shear wave velocity prediction model, the method further includes: If it is determined that the oil and gas well to be predicted is located in a different well area from the well that has been identified, then a correlation analysis is performed on the logging data and shear wave velocity data of the well that has been identified in the same well area as the oil and gas well to be predicted, and a second target logging data with a correlation higher than the preset value is obtained. If it is determined that the second target logging data and the first target logging data are of the same type, then the logging data of the oil and gas well to be predicted is input into the target shear wave velocity prediction model.
6. A transverse wave velocity prediction device for oil and gas wells, characterized in that, include: The acquisition module is used to acquire logging data and shear wave velocity data of identified oil and gas wells; The data analysis module classifies the logging data of the identified oil and gas wells according to data type, and obtains multiple sets of logging sequences; The correlation between each logging sequence and the shear wave velocity of the identified oil and gas wells is calculated, and the correlation between each logging sequence and the shear wave velocity of the identified oil and gas wells is output to obtain first target logging data with a correlation higher than a preset value. The first target logging data includes training data for model training, validation data for model hyperparameter optimization, and test data for obtaining model prediction errors. The correlation calculation between each logging sequence and the shear wave velocity of the identified oil and gas wells includes: use Perform correlation calculations; Where r represents the correlation between any set of logging sequences obtained from the classification and the shear wave velocity of the identified oil and gas wells, and n represents the number of sample points in the identified oil and gas wells. Let be the value of the logging sequence corresponding to the i-th sample point out of n sample points. This represents the average value of the logging sequences corresponding to n sample points. The shear wave velocity corresponding to the i-th sample point in the determined oil and gas well. Let be the mean of the shear wave velocities of all sample points in the determined oil and gas wells, where n is a natural number and i is a natural number greater than 0 and not greater than n; The model training module is used to input the training data into multiple different shear wave velocity prediction models for training, thereby obtaining multiple different initial shear wave velocity prediction models. The validation data is then input into each initial shear wave velocity prediction model for hyperparameter optimization, thereby obtaining a secondary shear wave velocity prediction model. The model training module is also used to input the test data into multiple different secondary shear wave velocity prediction models, obtain the shear wave velocity prediction value output by each secondary shear wave velocity prediction model, and output the shear wave velocity prediction model with the lowest shear wave velocity prediction value error among the multiple secondary shear wave velocity prediction models as the target shear wave velocity prediction model. The prediction module is used to input the logging data of the oil and gas well to be predicted into the target shear wave velocity prediction model and output the shear wave velocity of the oil and gas well to be predicted.
7. A device for predicting the shear wave velocity of oil and gas wells, characterized in that, include: At least one processor and memory; The memory is used to store computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1 to 5.
8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an instruction that a computer can execute, which, when executed by a processor, implements the method as described in any one of claims 1 to 5.