Method and device for predicting earthquake source fault plane parameters

Abstract

The present application relates to the field of artificial earthquake source simulation technology, and in particular to a method and device for predicting earthquake source fault plane parameters, wherein the method comprises: collecting a seismic motion inversion data set, pre-processing the parameters, establishing an original sample set containing input parameters and output parameters, and dividing the original sample set into training and test data sets according to a target ratio; using the training data set to train a mapping tilt random forest model to evaluate the importance of different input parameters and select the feature set with the smallest error as the optimal feature parameter set; using the optimal feature parameter set as a new input parameter, keeping the output parameters unchanged, and training the mapping tilt random forest model to obtain a final prediction model; using the test data set to verify the prediction effect of the final prediction model, and using the prediction model to predict the earthquake source fault plane parameters. Thus, the problem that the parameters of the earthquake source fault plane cannot be accurately estimated and the calculation accuracy and efficiency are difficult to be taken into account in the related art is solved.

Description

Technical Field

[0001] The present application relates to the technical field of artificial earthquake source simulation, and in particular to a method and device for predicting earthquake source fault plane parameters. Background Art

[0002] There are about 20 seismic belts in my country, and strong earthquakes are widely distributed. Therefore, it is very necessary to consider earthquake hazards in structural design.

[0003] Seismic input is the weakest link in structural seismic analysis. Seismic motions typically used in structural seismic design can be derived from actual seismic records of similar site conditions. However, due to the uneven distribution of existing strong earthquake records across regions, it is difficult to obtain actual records that meet engineering requirements. Therefore, manual simulation is required to generate seismic inputs suitable for different structures and site conditions.

[0004] Accurate estimation of focal fault plane parameters is fundamental to artificial ground motion simulation. However, related technologies primarily rely on linear regression for prediction, resulting in low accuracy. While the random forest algorithm, a commonly used machine learning method, possesses strong nonlinear fitting capabilities and robustness against noise, it struggles to balance computational accuracy and efficiency when dealing with multiple feature dimensions and large sample sizes, creating an urgent challenge. Summary of the Invention

[0005] The present application provides a method and device for predicting earthquake source fault plane parameters to solve the problems in related technologies such as the inability to accurately estimate earthquake source fault plane parameters and the difficulty in balancing calculation accuracy and efficiency.

[0006] The first aspect of the present application provides a method for predicting source fault plane parameters, including the following steps: collecting a seismic motion inversion data set, and preprocessing the parameters in the seismic motion inversion data set to establish an original sample set containing input parameters and output parameters, and dividing the original sample set into a training data set and a test data set according to a target ratio; using the training data set to train a mapping tilt random forest model to evaluate the importance of different input parameters and select a feature set with the smallest error as the optimal feature parameter set; using the optimal feature parameter set as a new input parameter and keeping the output parameter unchanged, training the mapping tilt random forest model to obtain a final prediction model; using the test data set to verify the final prediction model to verify the prediction effect of the final prediction model, and using the verified prediction model to predict source fault plane parameters.

[0007] Optionally, in one embodiment of the present application, the seismic motion inversion data set may include, but is not limited to, at least one of the earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, hypocenter longitude, hypocenter latitude, hypocenter depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, rupture velocity, and two-dimensional slip distribution of the fault plane.

[0008] Optionally, in one embodiment of the present application, the parameters in the seismic inversion dataset are preprocessed to establish an original sample set containing input parameters and output parameters, and the original sample set is divided into a training dataset and a test dataset according to a target ratio, including: summing the two-dimensional slip distribution of the fault plane along the strike direction and the dip direction respectively to obtain a one-dimensional distribution function of the slip in the corresponding direction; calculating the fault plane length and the fault plane width according to the one-dimensional distribution function of the slip in the corresponding direction to obtain the fault plane area; rescaling the slip distribution so that the seismic moment remains consistent with the previous one, The maximum value and average value of the scaled slip amount are used as the maximum slip of the fault plane and the average slip of the fault plane; the fault plane length, the fault plane width, the fault plane area, the maximum slip of the fault plane, and the average slip of the fault plane are classified as the output parameters; the earthquake date, the earthquake time, the moment magnitude, the Richter magnitude, the surface wave magnitude, the body wave magnitude, the focal longitude, the focal latitude, the focal depth, the fault plane strike, the fault plane dip, the fault plane slip angle, the slip type, the moment centroid depth, and the rupture velocity are classified as the input parameters.

[0009] Optionally, in one embodiment of the present application, the mapping tilted random forest model is trained using the training data set to evaluate the importance of different input parameters and select the feature set with the smallest error as the optimal feature parameter set, including: using the input parameters as features and the output parameters as labels to train each decision tree in the mapping tilted random forest model to obtain a trained mapping tilted random forest model; calculating the corresponding first out-of-bag data error for the out-of-bag data corresponding to each decision tree in the trained mapping tilted random forest model; using any feature of all sample data in the out-of-bag data corresponding to the decision tree as a test feature, and adding noise interference to the test feature to calculate the second out-of-bag data error corresponding to the decision tree; based on a preset importance formula, calculating the feature importance of the test feature according to the first out-of-bag data error and the second out-of-bag data error; and obtaining the optimal feature parameter set based on the feature importance of the test feature.

[0010] Optionally, in one embodiment of the present application, the calculation formula of the feature importance may be, but is not limited to:

[0011]

[0012] Among them, r1 and r2 are the out-of-bag errors before and after adding noise interference, and N is the number of decision trees in the mapping tilt random forest model.

[0013] The second aspect of the present application provides a device for predicting earthquake source fault surface parameters, including: an acquisition module for acquiring a seismic motion inversion data set, and preprocessing the parameters in the seismic motion inversion data set to establish an original sample set containing input parameters and output parameters, and dividing the original sample set into a training data set and a test data set according to a target ratio; an evaluation module for training a mapping tilt random forest model using the training data set to evaluate the importance of different input parameters and select a feature set with the smallest error as the optimal feature parameter set; a training module for using the optimal feature parameter set as a new input parameter and keeping the output parameter unchanged to train the mapping tilt random forest model to obtain a final prediction model; a prediction module for verifying the final prediction model using the test data set to verify the prediction effect of the final prediction model, and using the verified prediction model to predict earthquake source fault surface parameters.

[0014] Optionally, in one embodiment of the present application, the seismic motion inversion data set may include, but is not limited to, at least one of the earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, hypocenter longitude, hypocenter latitude, hypocenter depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, rupture velocity, and two-dimensional slip distribution of the fault plane.

[0015] Optionally, in one embodiment of the present application, the acquisition module includes: a summing unit for summing the two-dimensional slip distribution of the fault plane along the strike direction and the dip direction respectively to obtain a one-dimensional distribution function of the slip in the corresponding direction; a first calculation unit for calculating the length and width of the fault plane according to the one-dimensional distribution function of the slip in the corresponding direction to obtain the area of ​​the fault plane; a scaling unit for rescaling the slip distribution so that the seismic moment remains consistent with the previous one, and taking the maximum value and average value of the scaled slip as the maximum slip of the fault plane and the average slip of the fault plane. ; The first classification unit is used to classify the fault plane length, the fault plane width, the fault plane area, the maximum slip of the fault plane, and the average slip of the fault plane as the output parameters; the second classification unit is used to classify the earthquake date, the earthquake time, the moment magnitude, the Richter magnitude, the surface wave magnitude, the body wave magnitude, the focal longitude, the focal latitude, the focal depth, the fault plane strike, the fault plane dip, the fault plane slip angle, the slip type, the moment centroid depth, and the rupture velocity as the input parameters.

[0016] Optionally, in one embodiment of the present application, the evaluation module includes: a training unit, used to train each decision tree in the mapping tilted random forest model using the input parameters as features and the output parameters as labels to obtain a trained mapping tilted random forest model; a second calculation unit, used to calculate the corresponding first out-of-bag data error for the out-of-bag data corresponding to each decision tree in the trained mapping tilted random forest model; a third calculation unit, used to use any feature of all sample data in the out-of-bag data corresponding to the decision tree as a test feature, and add noise interference to the test feature to calculate the second out-of-bag data error corresponding to the decision tree; a fourth calculation unit, used to calculate the feature importance of the test feature based on the first out-of-bag data error and the second out-of-bag data error based on a preset importance formula; and an acquisition unit, used to obtain the optimal feature parameter set based on the feature importance of the test feature.

[0017] Optionally, in one embodiment of the present application, the calculation formula of the feature importance may be, but is not limited to:

[0018]

[0019] Among them, r1 and r2 are the out-of-bag errors before and after adding noise interference, and N is the number of decision trees in the mapping tilt random forest model.

[0020] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for predicting source fault plane parameters as described in the above embodiment.

[0021] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned method for predicting the parameters of the earthquake source fault plane.

[0022] The fifth aspect of the present application provides a computer program product, including a computer program, which is executed to implement the above-mentioned method for predicting earthquake source fault plane parameters.

[0023] The embodiment of the present application collects a seismic inversion data set to generate an original sample set, and divides it into training and test data sets. The training data set is used to train a mapping tilt random forest model to evaluate the importance of different input parameters, select the feature set with the smallest error as the optimal feature parameter set, use the optimal feature parameter set as the new input parameter, keep the output parameters unchanged, train the mapping tilt random forest model to obtain the final prediction model, and finally use the test data set to verify the prediction effect of the final prediction model, and use the prediction model to predict the source fault plane parameters, so that it can be used to predict the source fault plane parameters in engineering seismic analysis accurately, reliably and quickly. Therefore, it solves the problems in the related art that it is impossible to accurately estimate the source fault plane parameters and it is difficult to balance the calculation accuracy and efficiency.

[0024] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0026] Figure 1 This is a flow chart of a method for predicting earthquake source fault plane parameters according to an embodiment of the present application;

[0027] Figure 2 A schematic diagram of the importance of various characteristics of earthquake source fault plane parameters provided in one embodiment of the present application;

[0028] Figure 3 A schematic diagram of the fitting results of a training data set provided in one embodiment of the present application;

[0029] Figure 4 A schematic diagram of prediction results of a test data set provided for one embodiment of the present application;

[0030] Figure 5 A schematic diagram of the execution logic of a method for predicting earthquake source fault plane parameters provided by one embodiment of the present application;

[0031] Figure 6 Schematic diagram of the structure of a device for predicting earthquake source fault plane parameters according to an embodiment of the present application;

[0032] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0033] The following describes in detail embodiments of the present application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0034] The following describes the source fault plane parameter prediction method and device of the embodiment of the present application with reference to the accompanying drawings. In view of the problem that the related art mentioned in the above background technology cannot accurately estimate the source fault plane parameters and it is difficult to balance the calculation accuracy and calculation efficiency, the present application provides a source fault plane parameter prediction method, in which a seismic inversion data set can be collected to generate an original sample set, and the original sample set is divided into a training data set and a test data set. The training data set is used to train a mapping tilt random forest model to evaluate the importance of different input parameters, select the feature set with the smallest error as the optimal feature parameter set, use the optimal feature parameter set as the new input parameter, keep the output parameter unchanged, train the mapping tilt random forest model to obtain the final prediction model, and finally use the test data set to verify the prediction effect of the final prediction model, and use the prediction model to predict the source fault plane parameters, so that it can be used for the prediction of the source fault plane parameters in engineering seismic analysis, accurately, reliably and quickly. Thus, the problem that the related art cannot accurately estimate the source fault plane parameters and it is difficult to balance the calculation accuracy and calculation efficiency is solved.

[0035] Specifically, Figure 1 A flowchart of a method for predicting earthquake source fault plane parameters provided in an embodiment of the present application.

[0036] like Figure 1 As shown, the method for predicting earthquake source fault plane parameters includes the following steps:

[0037] In step S101 , a seismic inversion dataset is collected, and parameters in the seismic inversion dataset are preprocessed to establish an original sample set including input parameters and output parameters, and the original sample set is divided into a training dataset and a test dataset according to a target ratio.

[0038] It is understood that embodiments of the present application can collect a seismic inversion dataset, preprocess the parameters in the dataset, establish an original sample set containing input parameters and output parameters, and divide the original sample set into a training dataset and a test dataset according to a specified ratio. Furthermore, the ratio of the number of samples in the training dataset to the test dataset can be, but is not limited to, 1:1 to 9:1.

[0039] As an achievable method, the embodiment of the present application can collect a seismic inversion dataset. For example, the seismic inversion dataset can include, but is not limited to, 50 sets of data. The parameters can include, but are not limited to, moment magnitude, focal longitude, focal latitude, focal depth, fault plane strike, fault plane dip, fault plane slip angle, moment centroid depth, rupture velocity, and two-dimensional slip distribution on the fault plane. The embodiment of the present application can sum the two-dimensional slip distribution on the fault plane along the strike direction to obtain a one-dimensional distribution function f of the slip in this direction. s , and then calculate the fault plane length L by the following formula, which can be but not limited to:

[0040]

[0041] Where x is the position coordinate along the strike direction on the fault plane, and ACF is the calculated autocorrelation function.

[0042] In the embodiment of the present application, the fault plane length can be used as an output parameter, and the moment magnitude, focal longitude, focal latitude, focal depth, fault plane strike, fault plane dip, fault plane slip angle, moment centroid depth, and rupture velocity can be used as input parameters. The original sample set is shown in Table 1. Among them, the first 30 groups of data can be divided into training data sets, and the last 20 groups of data can be divided into test data sets. Therefore, the ratio of the number of samples in the training data set to the test data set is 3:2. Among them, Table 1 is the original sample set table, which can be expressed as follows:

[0043] Table 1

[0044]

[0045]

[0046] Therefore, the embodiment of the present application can collect a seismic inversion dataset and divide it into a training dataset and a test dataset, thereby providing data support for the subsequent training of a mapping tilt random forest model using the training dataset.

[0047] Optionally, in one embodiment of the present application, the seismic motion inversion data set may include, but is not limited to, at least one of the earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, hypocenter longitude, hypocenter latitude, hypocenter depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, rupture velocity, and two-dimensional slip distribution of the fault plane.

[0048] It should be noted that the seismic motion inversion dataset in the embodiments of the present application may include, but is not limited to, the following parameters: earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, focal longitude, focal latitude, focal depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, rupture velocity, and two-dimensional fault plane slip distribution. In actual applications, professionals skilled in the art may also add or filter these parameters based on actual conditions. This is merely an example and does not impose specific limitations.

[0049] Optionally, in one embodiment of the present application, the parameters in the seismic motion inversion data set are preprocessed to establish an original sample set containing input parameters and output parameters, and the original sample set is divided into a training data set and a test data set according to a target ratio, including: summing the two-dimensional slip distribution of the fault plane along the strike direction and the dip direction respectively to obtain a one-dimensional distribution function of the slip in the corresponding direction; calculating the fault plane length and the fault plane width according to the one-dimensional distribution function of the slip in the corresponding direction to obtain the fault plane area; rescaling the slip distribution so that the seismic moment is consistent with the previous one, and taking the maximum value and average value of the scaled slip as the maximum slip of the fault plane and the average slip of the fault plane; classifying the fault plane length, fault plane width, fault plane area, maximum slip of the fault plane, and average slip of the fault plane as output parameters; and classifying the earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, focal longitude, focal latitude, focal depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, and rupture velocity as input parameters.

[0050] It is understandable that in the embodiment of the present application, the two-dimensional slip distribution of the fault plane can be summed along the strike and dip directions respectively to obtain the one-dimensional distribution function f of the slip in these two directions. s and f d , and then calculate the fault plane length L and fault plane width W by the following formulas, which can be but not limited to:

[0051]

[0052]

[0053] Where x and y are the position coordinates along the strike and dip directions on the fault plane, respectively, and ACF is the calculated autocorrelation function.

[0054] Furthermore, the embodiment of the present application can also multiply the length and width of the fault plane to obtain the fault plane area, and rescale the slip distribution so that the seismic moment remains consistent with the previous one, and use the maximum and average values ​​of the scaled slip as the maximum slip of the fault plane and the average slip of the fault plane. The embodiment of the present application can classify the five parameters of fault plane length, fault plane width, fault plane area, maximum slip of the fault plane, and average slip of the fault plane as output parameters in the data set, and classify the 15 parameters of earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, focal longitude, focal latitude, focal depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, and rupture velocity as input parameters in the data set.

[0055] Therefore, the embodiment of the present application can calculate the output parameters and input parameters, thereby providing data support for subsequent importance evaluation of different input parameters to obtain the optimal feature parameter set.

[0056] In step S102, the training data set is used to train a mapping tilted random forest model to evaluate the importance of different input parameters and select the feature set with the smallest error as the optimal feature parameter set.

[0057] It can be understood that the embodiments of the present application can use the training data set to train the mapping tilt random forest model, evaluate the importance of different input parameters, and select the feature set with the smallest error as the optimal feature parameter set, thereby providing data support for subsequently using the optimal feature parameter set as the new input parameter to train the mapping tilt random forest model to obtain the final prediction model.

[0058] Optionally, in one embodiment of the present application, a mapping tilted random forest model can be trained using a training data set to evaluate the importance of different input parameters and select a feature set with the smallest error as the optimal feature parameter set, including: using the input parameters as features and the output parameters as labels to train each decision tree in the mapping tilted random forest model to obtain a trained mapping tilted random forest model; calculating the corresponding first out-of-bag data error for the out-of-bag data corresponding to each decision tree in the trained mapping tilted random forest model; using any feature of all sample data in the out-of-bag data corresponding to the decision tree as a test feature, and adding noise interference to the test feature to calculate the second out-of-bag data error corresponding to the decision tree; calculating the feature importance of the test feature based on the first out-of-bag data error and the second out-of-bag data error based on a preset importance formula; and obtaining the optimal feature parameter set based on the feature importance of the test feature.

[0059] It should be noted that the embodiment of the present application can use input parameters as features and output parameters as labels to train each decision tree in the mapping tilted random forest model. In the process of training each tree, the embodiment of the present application can, but is not limited to, use random sampling with replacement and randomly select feature subsets to reduce the correlation between decision trees, and calculate the error of the out-of-bag data corresponding to each decision tree trained in the mapping tilted random forest model. Subsequently, the embodiment of the present application can randomly use a certain feature of all sample data in the out-of-bag data corresponding to the above-mentioned decision tree as a test feature, add noise interference to the test feature, and then calculate the out-of-bag data error corresponding to the decision tree again. Finally, the embodiment of the present application can use a recursive feature elimination method to remove a specified proportion of features from the feature set in order from low to high importance, calculate the out-of-bag error rate before and after removal, and use the feature set with the smallest error rate as the optimal feature parameter set.

[0060] As an implementable method, the embodiment of the present application can use a training data set to train a mapping tilted random forest model. For example, the parameters of the mapping tilted random forest model can be set to: p = 9, d = 3, s = 3, n = 40. Then, the embodiment of the present application can use the input parameters as features and the output parameters as labels to train each decision tree in the mapping tilted random forest model. In the process of training each tree, random sampling with replacement and random selection of feature subsets are used to reduce the correlation between decision trees, and the out-of-bag data corresponding to each decision tree in the trained mapping tilted random forest model are calculated.

[0061] Then, a feature of all sample data in the out-of-bag data corresponding to the above decision tree is randomly selected as a test feature, and noise interference is added to the test feature. The out-of-bag data error corresponding to the decision tree is calculated again. Finally, the feature importance of each of the above test features is evaluated, and the importance is calculated according to the following formula, which can be but is not limited to:

[0062]

[0063] Among them, r1 and r2 are the out-of-bag errors before and after adding noise interference, and N is the number of decision trees in the mapping tilt random forest model.

[0064] Furthermore, in the embodiment of the present application, the importance indexes of all features can be summed and normalized to obtain the importance of each feature. Figure 2 This is a schematic diagram of the importance of each characteristic of the earthquake source fault plane parameter provided in one embodiment of the present application. Figure 2 As shown in the figure, the higher the importance, the greater the correlation between these parameters and the results.

[0065] Next, the embodiment of the present application can use a recursive feature elimination method to remove a specified proportion of features from the feature set in order of importance from low to high, calculate the out-of-bag error rate before and after removal, and obtain the feature set with the smallest error rate: moment magnitude, focal depth, fault plane inclination, fault plane slip angle, moment centroid depth, and rupture velocity, which is used as the optimal feature parameter set.

[0066] Therefore, the embodiment of the present application obtains the optimal feature parameter set through calculation, thereby providing data support for subsequently using it as a new input parameter to train the mapping tilted random forest model to obtain the final prediction model.

[0067] Optionally, in one embodiment of the present application, the calculation formula for feature importance may be, but is not limited to,:

[0068]

[0069] Among them, r1 and r2 are the out-of-bag errors before and after adding noise interference, and N is the number of decision trees in the mapping tilt random forest model.

[0070] It should be noted that the embodiment of the present application includes the out-of-bag errors r1 and r2 before and after adding noise interference and the number of decision trees N in the mapping tilt random forest model, and the value of feature importance is calculated by the following formula, which can be but is not limited to:

[0071]

[0072] Furthermore, in the embodiment of the present application, the importance indexes of all features can be summed and normalized to obtain the importance of each feature, such as Figure 2 As shown, the evaluation indicators can be selected from mean absolute error, mean square error, root mean square error, determination coefficient, quantile loss, and Huber loss.

[0073] Therefore, the embodiment of the present application can obtain the calculation formula of feature importance by calculating the out-of-bag errors r1 and r2 before and after noise interference and the number of decision trees N in the mapping tilt random forest model, thereby providing a theoretical basis for the subsequent calculation of the optimal feature parameter set to train the mapping tilt random forest model.

[0074] In step S103, the optimal feature parameter set is used as a new input parameter, and the output parameter is kept unchanged, and the mapping tilt random forest model is trained to obtain the final prediction model.

[0075] It can be understood that after obtaining the optimal feature parameter set based on the feature importance of the test features, the embodiment of the present application can further use the optimal feature parameter set as a new input parameter and keep the output parameter unchanged to train the mapping tilted random forest model to obtain the final prediction model.

[0076] It should be noted that, in the embodiment of the present application, moment magnitude, focal depth, fault plane dip, fault plane slip angle, centroid depth, and rupture velocity are used as input parameters, and fault plane length is used as output parameter. The parameters of the mapping tilt random forest model are set to: p = 6, d = 2, s = 3, n = 40, and the final prediction model is obtained by training. Figure 3 As shown, the present application embodiment provides a schematic diagram of the training data set fitting results. Figure 3 It can be seen that the fitting value of the embodiment of the present application is very close to the true value, and the fitting effect is good.

[0077] As a feasible method, the embodiment of the present application can generate a corresponding tilted mapping matrix Q for each decision tree in the mapping tilted random forest model. The matrix size is p*d, and it is composed of randomly distributed -1, 0, and 1. The probability of occurrence of the three values ​​is 1 / 2s, 1-1 / s, and 1 / 2s, respectively, where p is the number of input features; d is the number of retained features, and the value range of d is 1 to p; s is the sparsity, and the value range of s is 2 to positive infinity. Furthermore, the embodiment of the present application regards the training data as a training matrix of size n*p, where n is the amount of data and p is the number of features. The training matrix is ​​multiplied by the tilted mapping matrix to obtain the sparse matrix required for training each decision tree. Finally, the embodiment of the present application uses the sparse matrix to train each decision tree to obtain the final mapping tilted random forest model.

[0078] Therefore, the embodiment of the present application trains the mapping tilt random forest model to obtain the final prediction model, so that the mapping tilt random forest model has strong fitting ability and good adaptability to multi-feature input. It is used to predict the parameters of the earthquake source fault plane, and the prediction results are more accurate and reliable than traditional methods.

[0079] In step S104, the final prediction model is verified using the test data set to verify the prediction effect of the final prediction model, and the verified prediction model is used to predict the source fault plane parameters.

[0080] It is understandable that the embodiment of the present application can use the test data set to verify the final mapping tilt random forest model to verify the prediction effect of the final prediction model, and use the verified prediction model to predict the source fault plane parameters.

[0081] As an achievable method, the embodiment of the present application can use the final prediction model and the test data set for verification to verify the prediction effect of the mapping tilt random forest model. Figure 4 As shown, the embodiment of the present application provides a schematic diagram of the prediction results of the test data set. Figure 4It can be seen that the predicted value is close to the true value, indicating that the model prediction effect is good. In order to further illustrate the superiority of the mapping tilt random forest model used in the embodiment of this application, traditional random forest, support vector machine, and neural network can be selected for modeling and comparative analysis. The root mean square error and determination coefficient are used to quantitatively measure the model prediction accuracy. Table 2 is a model prediction accuracy table, which can be expressed as follows:

[0082] Table 2

[0083] Root mean square error Coefficient of determination Mapping skewed random forests 10.48 0.98 Traditional Random Forest 20.98 0.92 Support Vector Machine 26.96 0.88 BP neural network 29.95 0.86

[0084] From the results, it can be seen that the root mean square error of the prediction results of the mapping tilt random forest model is smaller and the determination coefficient is closest to 1, indicating that the prediction accuracy of the mapping tilt random forest model is better than that of the simple traditional random forest, support vector machine, and BP neural network, and it has a good application effect in the source fault plane parameters.

[0085] Therefore, the prediction model verified by the embodiment of the present application can provide reliable prediction of source fault plane parameters for studies such as probabilistic earthquake hazard analysis, regional seismic motion synthesis, and maximum credible seismic motion simulation.

[0086] The following describes a specific embodiment and a schematic diagram of the execution logic of the method for predicting source fault plane parameters of the embodiment of the present application in combination with the accompanying drawings.

[0087] Figure 5 This is a schematic diagram of the execution logic of the earthquake source fault plane parameter prediction method.

[0088] like Figure 5 As shown, the prediction logic of the source fault plane parameters of this application is as follows:

[0089] S501, collect the seismic inversion data set, and execute S502;

[0090] S502: Use the data set as the original sample set and execute S503 and S504;

[0091] S503, training data set, executing S505;

[0092] S504, test data set, execute S508;

[0093] S505: Parameter importance evaluation, and then proceed to S506;

[0094] S506, calculate the optimal feature parameter set, and execute S507;

[0095] S507, perform final model prediction, and execute S508;

[0096] S508: Verify the prediction effect.

[0097] According to a method for predicting source fault plane parameters proposed in an embodiment of the present application, in this method, a seismic inversion data set can be collected to generate an original sample set, which is divided into training and test data sets. The training data set is used to train a mapping tilt random forest model to evaluate the importance of different input parameters, select the feature set with the smallest error as the optimal feature parameter set, use the optimal feature parameter set as the new input parameter, keep the output parameters unchanged, train the mapping tilt random forest model to obtain the final prediction model, and finally use the test data set to verify the prediction effect of the final prediction model, and use the prediction model to predict the source fault plane parameters, so that it can be used to predict the source fault plane parameters in engineering seismic analysis accurately, reliably and quickly. As a result, the problems in the related art such as the inability to accurately estimate the source fault plane parameters and the difficulty in balancing calculation accuracy and efficiency are solved.

[0098] Next, the earthquake source fault plane parameter prediction device proposed in accordance with the embodiment of the present application will be described with reference to the accompanying drawings.

[0099] Figure 6 It is a structural diagram of the earthquake source fault plane parameter prediction device of an embodiment of the present application.

[0100] like Figure 6 As shown, the earthquake source fault plane parameter prediction device 10 includes: an acquisition module 100, an evaluation module 200, a training module 300 and a prediction module 400.

[0101] Among them, the acquisition module 100 is used to acquire the seismic inversion dataset and preprocess the parameters in the seismic inversion dataset to establish an original sample set containing input parameters and output parameters, and divide the original sample set into a training dataset and a test dataset according to a target ratio.

[0102] The evaluation module 200 is used to train the mapping tilt random forest model using the training data set to evaluate the importance of different input parameters and select the feature set with the smallest error as the optimal feature parameter set.

[0103] The training module 300 is used to train the mapping tilt random forest model by taking the optimal feature parameter set as a new input parameter and keeping the output parameter unchanged to obtain the final prediction model.

[0104] The prediction module 400 is used to verify the final prediction model using the test data set to verify the prediction effect of the final prediction model, and use the verified prediction model to predict the source fault plane parameters.

[0105] Optionally, in one embodiment of the present application, the seismic motion inversion data set may include, but is not limited to, at least one of the earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, hypocenter longitude, hypocenter latitude, hypocenter depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, rupture velocity, and two-dimensional slip distribution of the fault plane.

[0106] Optionally, in one embodiment of the present application, the acquisition module 100 includes: a summing unit, a first calculating unit, a scaling unit, a first classification unit, and a second classification unit.

[0107] The summing unit is used to sum the two-dimensional slip distribution of the fault plane along the strike direction and the dip direction respectively to obtain a one-dimensional distribution function of the slip in the corresponding direction.

[0108] The first calculation unit is used to calculate the fault plane length and the fault plane width according to the one-dimensional distribution function of the slip amount in the corresponding direction to obtain the fault plane area.

[0109] The scaling unit is used to rescale the slip distribution so that the seismic moment is consistent with the previous one, and the maximum and average values ​​of the scaled slip are used as the maximum slip and average slip of the fault plane.

[0110] The first classification unit is used to classify the fault plane length, fault plane width, fault plane area, maximum slip of the fault plane, and average slip of the fault plane as output parameters.

[0111] The second classification unit is used to classify the earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, focal longitude, focal latitude, focal depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, and rupture velocity as input parameters.

[0112] Optionally, in one embodiment of the present application, the evaluation module 200 includes: a training unit, a second calculation unit, a third calculation unit, a fourth calculation unit and an acquisition unit.

[0113] The training unit is used to train each decision tree in the mapping tilted random forest model using the input parameters as features and the output parameters as labels to obtain a trained mapping tilted random forest model.

[0114] The second calculation unit is used to calculate the corresponding first out-of-bag data error for the out-of-bag data corresponding to each decision tree in the trained mapping tilt random forest model.

[0115] The third calculation unit is used to use any feature of all sample data in the out-of-bag data corresponding to the decision tree as a test feature, and add noise interference to the test feature to calculate the second out-of-bag data error corresponding to the decision tree.

[0116] The fourth calculation unit is used to calculate the feature importance of the test feature according to the first out-of-bag data error and the second out-of-bag data error based on a preset importance formula.

[0117] The acquisition unit is used to obtain an optimal feature parameter set based on the feature importance of the test feature.

[0118] Optionally, in one embodiment of the present application, the calculation formula of feature importance may be, but is not limited to:

[0119]

[0120] Among them, r1 and r2 are the out-of-bag errors before and after adding noise interference, and N is the number of decision trees in the mapping tilt random forest model.

[0121] It should be noted that the above explanation of the embodiment of the method for predicting the parameters of the earthquake source fault plane is also applicable to the device for predicting the parameters of the earthquake source fault plane of this embodiment, and will not be repeated here.

[0122] According to the earthquake source fault plane parameter prediction device proposed in the embodiment of the present application, first, the seismic inversion data set is collected by the acquisition module to establish an original sample set, and the original sample set is divided into a training data set and a test data set; secondly, the evaluation module trains the mapping tilt random forest model to evaluate the importance of different input parameters and selects the feature set with the smallest error as the optimal feature parameter set; furthermore, the training module uses the optimal feature parameter set as the new input parameter and keeps the output parameter unchanged to train the mapping tilt random forest model to obtain the final prediction model; finally, the final prediction model is verified by the prediction module to verify the prediction effect of the final prediction model, and the prediction model after verification is used to predict the earthquake source fault plane parameters, so that it can be used for the prediction of the earthquake source fault plane parameters in engineering seismic analysis, accurately, reliably and quickly. Thus, the problems in the related art that the earthquake source fault plane parameters cannot be accurately estimated and the calculation accuracy and efficiency are difficult to be taken into account are solved.

[0123] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0124] A memory 701 , a processor 702 , and a computer program stored in the memory 701 and executable on the processor 702 .

[0125] When the processor 702 executes the program, the method for predicting the parameters of the earthquake source fault plane provided in the above embodiment is implemented.

[0126] Furthermore, the electronic device further includes:

[0127] The communication interface 703 is used for communication between the memory 701 and the processor 702 .

[0128] The memory 701 is used to store computer programs that can be run on the processor 702 .

[0129] The memory 701 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0130] If the memory 701, processor 702, and communication interface 703 are implemented independently, the communication interface 703, memory 701, and processor 702 can be interconnected via a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 7 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0131] Optionally, in a specific implementation, if the memory 701, the processor 702 and the communication interface 703 are integrated on a chip, the memory 701, the processor 702 and the communication interface 703 can communicate with each other through an internal interface.

[0132] The processor 702 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0133] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for predicting earthquake source fault plane parameters.

[0134] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned method for predicting earthquake source fault plane parameters.

[0135] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0136] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this application, "N" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0137] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing a custom logical function or process step, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed in a different order than shown or discussed, including performing functions in a substantially simultaneous manner or in a reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application pertain.

[0138] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or N wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program can be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing it in other suitable ways as necessary, and then storing it in a computer memory.

[0139] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiment, the N steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented using hardware, as in another embodiment, it can be implemented using any one or a combination of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0140] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0141] In addition, the functional units in the various embodiments of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into a module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0142] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A method for predicting earthquake source fault plane parameters, characterized in that: The following steps are involved: Acquiring a seismic inversion dataset, and preprocessing parameters in the seismic inversion dataset to establish an original sample set including input parameters and output parameters, and dividing the original sample set into a training dataset and a test dataset according to a target ratio; Using the training data set to train a mapping tilted random forest model to evaluate the importance of different input parameters, and select the feature set with the smallest error as the optimal feature parameter set; Using the optimal feature parameter set as a new input parameter and keeping the output parameter unchanged, training the mapping tilt random forest model to obtain a final prediction model; The final prediction model is verified using the test data set to verify the prediction effect of the final prediction model, and the verified prediction model is used to predict the source fault plane parameters.

2. The method according to claim 1, characterized in that The seismic motion inversion data set includes at least one of the earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, focal longitude, focal latitude, focal depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, rupture velocity, and two-dimensional slip distribution of the fault plane.

3. The method according to claim 2, characterized in that The preprocessing of the parameters in the seismic inversion dataset to establish an original sample set including input parameters and output parameters, and dividing the original sample set into a training dataset and a test dataset according to a target ratio, includes: The two-dimensional slip distribution of the fault plane is summed along the strike direction and the dip direction respectively to obtain a one-dimensional distribution function of the slip in the corresponding direction; Calculating the fault plane length and the fault plane width according to the one-dimensional distribution function of the slip amount in the corresponding direction to obtain the fault plane area; The slip distribution is rescaled to make the seismic moment consistent with the previous one, and the maximum and average values ​​of the rescaled slip are used as the maximum slip and average slip of the fault plane; Classifying the fault plane length, the fault plane width, the fault plane area, the maximum slip of the fault plane, and the average slip of the fault plane as the output parameters; The earthquake date, the earthquake time, the moment magnitude, the Richter magnitude, the surface wave magnitude, the body wave magnitude, the hypocenter longitude, the hypocenter latitude, the hypocenter depth, the fault plane strike, the fault plane dip, the fault plane slip angle, the slip type, the moment centroid depth, and the rupture velocity are classified as the input parameters.

4. The method according to claim 1, wherein The mapping tilt random forest model is trained using the training data set to evaluate the importance of different input parameters and select the feature set with the smallest error as the optimal feature parameter set, including: Training each decision tree in the mapping-skewed random forest model using the input parameters as features and the output parameters as labels to obtain a trained mapping-skewed random forest model; Calculating a first out-of-bag data error for the out-of-bag data corresponding to each decision tree in the trained mapped tilted random forest model; Taking any feature of all sample data in the out-of-bag data corresponding to the decision tree as a test feature, and adding noise interference to the test feature to calculate the second out-of-bag data error corresponding to the decision tree; Calculating the feature importance of the test feature according to the first out-of-bag data error and the second out-of-bag data error based on a preset importance formula; The optimal feature parameter set is obtained based on the feature importance of the test features.

5. The method according to claim 4, characterized in that The calculation formula for the feature importance is: Among them, r1 and r2 are the out-of-bag errors before and after adding noise interference, and N is the number of decision trees in the mapping tilt random forest model.

6. A device for predicting earthquake source fault plane parameters, characterized in that: include: an acquisition module, configured to acquire a seismic inversion dataset, preprocess parameters in the seismic inversion dataset to establish an original sample set including input parameters and output parameters, and divide the original sample set into a training dataset and a test dataset according to a target ratio; An evaluation module is used to train a mapping tilted random forest model using the training data set to evaluate the importance of different input parameters and select a feature set with the smallest error as the optimal feature parameter set; A training module is used to train a mapping tilt random forest model using the optimal feature parameter set as a new input parameter and keeping the output parameter unchanged to obtain a final prediction model; The prediction module is used to verify the final prediction model using the test data set to verify the prediction effect of the final prediction model, and use the verified prediction model to predict the source fault plane parameters.

7. The device according to claim 6, characterized in that The seismic motion inversion data set includes at least one of the earthquake date, earthquake time, moment magnitude, Richter magnitude, surface wave magnitude, body wave magnitude, focal longitude, focal latitude, focal depth, fault plane strike, fault plane dip, fault plane slip angle, slip type, moment centroid depth, rupture velocity, and two-dimensional slip distribution of the fault plane.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for predicting source fault plane parameters according to any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method for predicting source fault plane parameters as described in any one of claims 1 to 5.

10. A computer program product comprising a computer program, characterized in that The computer program is executed to implement the method for predicting source fault plane parameters according to any one of claims 1 to 5.

Images