Earthquake sand liquefaction deformation prediction method based on machine learning

Abstract

The invention discloses an earthquake sand liquefaction deformation prediction method based on machine learning, and relates to the technical field of earthquake engineering and machine learning. Missing value processing, abnormal value detection, feature selection and standardization processing are carried out on the multi-source coupling seismic data set to obtain a seismic sand liquefaction deformation data set, a seismic sand liquefaction deformation prediction model is constructed according to the seismic sand liquefaction deformation data set, the root-mean-square error, the mean absolute error and the decision coefficient of the model are calculated, and the seismic sand liquefaction deformation prediction result is obtained. And according to corresponding threshold judgment, carrying out hyper-parameter adjustment on the model through Bayesian optimization or carrying out model packaging through Flask, processing the obtained new seismic sand liquefaction characteristic data, and inputting the processed data into the packaged model to obtain a seismic sand liquefaction deformation prediction result. The method can effectively predict earthquake sand liquefaction deformation, and provides support for earthquake disaster prevention and engineering design.

Description

TECHNICAL FIELD

[0001] The present application relates to the fields of machine learning and geotechnical engineering, and in particular to a method for predicting seismic sand liquefaction deformation based on machine learning. BACKGROUND

[0002] In the field of earthquake engineering and geotechnical engineering, the phenomenon of seismic sand liquefaction poses a great threat to the safety of building facilities and geological structures. Traditional methods for predicting sand liquefaction are mainly based on empirical formulas and simple physical models, relying on limited field data and theoretical assumptions, and are difficult to fully consider the complex geological conditions, seismic characteristics, and the interaction between the two. In the face of complex and variable actual situations, the prediction accuracy of traditional methods is limited, and their applicability is also limited.

[0003] With the rapid development of machine learning technology, machine learning technology has shown great predictive and analytical capabilities in many fields. In the field of earthquake engineering, machine learning provides new ideas and methods for predicting seismic sand liquefaction deformation. However, existing methods still have problems. Seismic sand liquefaction is influenced by multiple factors such as groundwater level, soil distribution, and seismic parameters. Existing methods lack effective integration of multi-source information, resulting in insufficient completeness and representativeness of the data set. The processing of missing values and outliers and the feature selection method are relatively simple, which can easily miss key features or introduce noise, affecting the model training effect. At the same time, the existing method relies on empirical settings for hyperparameter adjustment of the machine learning model, making it difficult to automatically search for the optimal parameter combination, and the model deployment method is single, lacking a convenient interface to support real-time prediction and cross-system application. Therefore, how to effectively integrate multi-source data, optimize the data processing process, improve the model performance, and facilitate practical application of machine learning in predicting seismic sand liquefaction deformation has become a problem to be solved. Therefore, a method for predicting seismic sand liquefaction deformation based on machine learning is proposed. SUMMARY

[0004] The present application aims to provide a method for predicting seismic sand liquefaction deformation based on machine learning to solve the problems raised in the background.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions: A method for predicting seismic sand liquefaction deformation based on machine learning, comprising the following steps: S1. Obtain the seismic sand liquefaction characteristic data and the sand liquefaction deformation of the seismic case from the engineering investigation report and the monitoring data of the historical seismic case, and construct a multi-source coupled seismic data set accordingly; S2. The missing value processing is performed on the multi-source coupled seismic data set, the multi-source coupled seismic data set after the missing value processing is subjected to abnormality detection by using IQR, the multi-source coupled seismic data set after the abnormality detection is subjected to feature selection processing, meanwhile, the Z-score standardization is used for standardization processing, the seismic sandy soil liquefaction deformation data set is obtained, and is stored into a database; S3. According to the seismic sandy soil liquefaction deformation data set, a seismic sandy soil liquefaction deformation prediction model is constructed, the root mean square error, the mean absolute error and the determination coefficient of the seismic sandy soil liquefaction deformation prediction model are calculated, the constructed seismic sandy soil liquefaction deformation prediction model is judged according to the root mean square error threshold, the mean absolute error threshold and the determination coefficient threshold, and a judgment result is output; S4. According to the judgment result, the seismic sandy soil liquefaction deformation prediction model is subjected to hyperparameter adjustment or model packaging, the seismic sandy soil liquefaction deformation prediction model after the hyperparameter adjustment is retrained and updated through the seismic sandy soil liquefaction deformation data set; S5. New seismic sandy soil liquefaction feature data is acquired, the new seismic sandy soil liquefaction feature data is subjected to missing value processing, abnormal value detection, feature selection processing and standardization processing, and is input into the packaged seismic sandy soil liquefaction deformation prediction model, and a seismic sandy soil liquefaction deformation prediction result is obtained.

[0006] Preferably, the method for constructing the multi-source coupled seismic data set comprises the following steps: According to the engineering investigation report and the detection data of the historical earthquake case, the groundwater depth, the soil layer burial depth, the soil layer distribution, the relative density, the water content, the effective overburden stress, the clay content, the standard penetration number, the soil particle specific gravity, the void ratio, the peak ground acceleration, the earthquake duration, the predominant period, the velocity response spectrum value, the magnitude, the epicentral distance and the ground settlement of the earthquake case are obtained, the groundwater depth, the soil layer burial depth and the soil layer distribution reflect the geological conditions of the earthquake case occurrence area, the relative density, the water content, the effective overburden stress, the clay content, the standard penetration number, the soil particle specific gravity and the void ratio reflect the physical and mechanical properties of the soil body in the earthquake case occurrence area, and the peak ground acceleration, the earthquake duration, the predominant period, the velocity response spectrum value, the magnitude and the epicentral distance reflect the earthquake intensity and characteristics of the earthquake case; The geographical position information and the unique ID number of each earthquake case are manually labeled, the earthquake sandy soil liquefaction characteristic data and the sandy soil liquefaction deformation corresponding to the earthquake case are matched with the geographical position information and the unique ID number of the earthquake case according to the geographical position information and the unique ID number of the earthquake case, the correlation is established, the multi-source coupled earthquake data set is constructed, the multi-source coupled earthquake data set is composed of the geographical position information and the unique ID number of the earthquake case and the earthquake sandy soil liquefaction characteristic data and the sandy soil liquefaction deformation determined by the geographical position information and the unique ID number of the earthquake case, the earthquake sandy soil liquefaction characteristic data includes the groundwater depth, the soil layer depth, the soil layer distribution, the relative density, the water content, the effective overburden stress, the clay content, the standard penetration number, the soil particle specific gravity, the void ratio, the peak ground acceleration, the earthquake duration, the predominant period, the velocity response spectrum value, the magnitude and the epicentral distance, and the sandy soil liquefaction deformation is the ground settlement, and the ground settlement is used for measuring the deformation degree of the earthquake sandy soil liquefaction.

[0007] Preferably, the method for performing missing value processing and anomaly detection on the multi-source coupled earthquake data set and standardizing the multi-source coupled earthquake data set after feature selection processing comprises the following steps: The mean filling method is used to fill the missing values by using the mean filling method formula, the missing values are filled by calculating the average value of the existing data, the overall characteristics of the data can be maintained to a certain extent, and data deviation caused by the missing values is avoided; The mean filling method formula is: wherein, is the result obtained after the mean filling method processing and used for filling the missing values, is the weight of each data when calculating the average value, is the operation of summing all the data in the data set S, is the ith data in the data set S; For the multi-source coupled earthquake data set after the missing value processing, the first quartile Q1 and the third quartile Q3 of the data set are calculated, IQR = Q3-Q1 is obtained, a reasonable range is determined, generally, the data points less than Q1-1.5*IQR or greater than Q3+1.5*IQR are regarded as abnormal values, once the abnormal values are identified in the multi-source coupled earthquake data set after the missing value processing, the abnormal data are discarded, the interference of the abnormal data on the subsequent analysis and model training is avoided, and the reliability and accuracy of the data are ensured; For the multi-source coupled earthquake data set after the feature selection processing, the Z-score standardization formula is used for standardization, after the standardization processing, the mean value of the data in the data set becomes 0, the standard deviation becomes 1, and all the data are mapped to a distribution with a unified scale; The Z-score standardization formula is: ; wherein, is the original data value, is the mean of the data, is the standard deviation of the data.

[0008] Preferably, the method for performing feature selection processing on the multi-source coupled seismic data set after anomaly detection and obtaining a seismic sand liquefaction deformation data set comprises:

[0009] calculating the Spearman correlation coefficient of all feature elements in the multi-source coupled seismic data set after anomaly detection by a Spearman correlation coefficient formula , setting the correlation threshold value to 0.1, comparing the absolute value of the Spearman correlation coefficient of all feature elements with the correlation threshold value in size, deleting the feature elements with a Spearman correlation coefficient less than or equal to the correlation threshold value from the multi-source coupled seismic data set, thereby realizing feature selection processing on the multi-source coupled seismic data set;

[0010] The Spearman correlation coefficient formula is: wherein, is the Spearman correlation coefficient of the feature element, is the rank difference of the feature element and the i-th observation value of the sand liquefaction deformation amount, is the sample size; performing standardization processing on the multi-source coupled seismic data set subjected to feature selection processing to obtain a seismic sand liquefaction deformation data set, wherein the seismic sand liquefaction deformation data set comprises a sand liquefaction deformation amount and seismic sand liquefaction characteristic data corresponding to the sand liquefaction deformation amount subjected to feature selection processing. The feature element is a groundwater depth, a soil layer burial depth, a soil layer distribution, a relative density, a water content, an effective overburden stress, a clay content, a standard penetration number, a soil particle specific gravity, a void ratio, a peak ground acceleration, a seismic duration, a predominant period, a velocity response spectrum value, a magnitude, or an epicentral distance.

[0011] Preferably, the method for constructing a seismic sand liquefaction deformation prediction model according to the seismic sand liquefaction deformation data set comprises: The seismic sandy soil liquefaction deformation data set is obtained from a database, the seismic sandy soil liquefaction deformation data set is divided into a training set and a test set according to a ratio of 8:2, an XGBoost model is constructed by using an xgboost library, and hyperparameters of the XGBoost model, including a number of trees, a learning rate, a maximum depth of a tree, a subsample ratio and a column sample ratio, are set; the xgboost library is a gradient boosting library, and the XGBoost model is a machine learning algorithm; The XGBoost model is trained by using the training set, and a trained XGBoost model, that is, a seismic sandy soil liquefaction deformation prediction model, is obtained.

[0012] Preferably, the method for judging the constructed seismic sandy soil liquefaction deformation prediction model according to the root mean square error threshold, the mean absolute error threshold and the coefficient of determination threshold, and outputting a judgment result is as follows: The test set is input into the seismic sandy soil liquefaction deformation prediction model, the model predicts the seismic sandy soil liquefaction deformation according to the input data and the rules learned in the model training process, and outputs a prediction result, that is, a predicted sandy soil liquefaction deformation According to the predicted sandy soil liquefaction deformation and the real sandy soil liquefaction deformation in the test set, the root mean square error RMSE, the mean absolute error MAE and the coefficient of determination of the seismic sandy soil liquefaction deformation prediction model are obtained by using a root mean square error formula, a mean absolute error formula and a coefficient of determination formula respectively. ; The root mean square error formula RMSE and the mean absolute error formula MAE are as follows: Wherein, N is the number of samples, is the predicted sandy soil liquefaction deformation of the i th sample, is the actual sandy soil liquefaction deformation of the i th sample, and are the root mean square error and the mean absolute error of the seismic sandy soil liquefaction deformation prediction model respectively. The coefficient of determination formula is as follows: Wherein, is the number of samples, is the predicted sandy soil liquefaction deformation of the i th sample, is the actual sandy soil liquefaction deformation of the i th sample, is the coefficient of determination of the seismic sandy soil liquefaction deformation prediction model, is the mean value of the real sandy soil liquefaction deformation of all samples. The root mean square error, the mean absolute error and the determination coefficient of the seismic sandy soil liquefaction deformation prediction model are compared with the root mean square error threshold value, the mean absolute error threshold value and the determination coefficient threshold value respectively. If the root mean square error of the seismic sandy soil liquefaction deformation prediction model is less than the root mean square error threshold value 1, the mean absolute error is less than the mean absolute error threshold value 0.8, and the determination coefficient is greater than the determination coefficient threshold value 0.7, the judgment result is that the seismic sandy soil liquefaction deformation prediction model is deployed. Otherwise, the seismic sandy soil liquefaction deformation prediction model is adjusted.

[0013] Preferably, the method for adjusting the hyperparameters of the seismic sandy soil liquefaction deformation prediction model comprises: The hyperparameters of the seismic sandy soil liquefaction deformation prediction model are adjusted through Bayesian optimization, which is a hyperparameter tuning method based on the Bayes theorem and can efficiently search for the optimal combination of hyperparameters in the hyperparameter search space. The hyperparameter search space is a set of value ranges set for the hyperparameters in the machine learning model tuning process.

[0014] Preferably, the method for packaging the seismic sandy soil liquefaction deformation prediction model comprises: The seismic sandy soil liquefaction deformation prediction model is packaged into a Web service interface through Flask, so as to realize the packaging of the seismic sandy soil liquefaction deformation prediction model.

[0015] Preferably, the method for obtaining the seismic sandy soil liquefaction deformation prediction result comprises: The new seismic sandy soil liquefaction feature data processed by the missing value processing, the abnormal value detection, the feature selection processing and the standardization processing are packaged into a data packet in the JSON format, the data packet is sent to the application programming interface of the server through the HTTP protocol, the server obtains the data packet through the application programming interface, calls the seismic sandy soil liquefaction deformation prediction model packaged into the Web service interface, the seismic sandy soil liquefaction deformation prediction model outputs the seismic sandy soil liquefaction deformation prediction result according to the data packet, and returns the seismic sandy soil liquefaction deformation prediction result through the HTTP protocol.

[0016] Compared with the prior art, the present application has the following beneficial effects: 1.The application realizes the construction of multi-source data integration and high-quality data processing system, obtains the multi-source coupling characteristic data of geological conditions, soil physical and mechanical properties and seismic characteristics through the engineering investigation report and monitoring data of historical earthquake cases, establishes the correlation through geographic location information and unique ID number, solves the problem of insufficient data set integrity in the prior art, constructs a more representative multi-source coupling seismic data set, adopts the mean filling method to process missing values, the IQR method to remove outliers, the Spearman correlation coefficient to screen key features and the Z-score standardization to form a full-process data processing system of missing value repair, outlier filtering, feature selection and scale unification, effectively retains the core features and filters the noise, significantly improves the quality of input model data, and lays a data foundation for high-precision prediction.

[0017] 2.The application realizes intelligent model optimization and convenient engineering application support, constructs a prediction model based on the XGBoost algorithm, introduces the Bayesian optimization technology to automatically optimize the hyperparameters of the model, breaks through the limitations of relying on experience setting in the prior art, efficiently searches for the optimal parameter combination, improves the adaptability and prediction accuracy of the model to complex geology and seismic conditions, encapsulates the model into a standardized Web service interface through the lightweight Web framework Flask, supports JSON data interaction based on the HTTP protocol, realizes cross-system calling and real-time prediction, solves the problem of single deployment and interface missing in traditional methods, is convenient for integration with the earthquake monitoring and engineering design platform, at the same time, establishes a multi-dimensional model evaluation mechanism, forms a closed loop of training, evaluation, optimization and deployment, ensures continuous optimization of the model performance, and provides reliable technical support for earthquake disaster prevention and infrastructure design. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of these drawings.

[0019] Figure 1 The step flow chart of the present application. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical scheme and advantages of the present application more clear, the technical scheme of the present application will be described in detail below. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0021] Embodiments, such as Figure 1 A machine learning-based seismic sand liquefaction deformation prediction method, as shown in the embodiments, includes the following steps: S1. Obtain seismic sand liquefaction characteristic data and sand liquefaction deformation of a seismic case through engineering investigation reports and monitoring data of historical seismic cases, and construct a multi-source coupled seismic data set accordingly; S2. Perform missing value processing on the multi-source coupled seismic data set, perform anomaly detection on the multi-source coupled seismic data set after missing value processing using IQR, perform feature selection processing on the multi-source coupled seismic data set after anomaly detection, and simultaneously perform standardization processing using Z-score standardization to obtain a seismic sand liquefaction deformation data set, and store it in a database; S3. Construct a seismic sand liquefaction deformation prediction model according to the seismic sand liquefaction deformation data set, calculate the root mean square error, the mean absolute error, and the determination coefficient of the seismic sand liquefaction deformation prediction model, judge the constructed seismic sand liquefaction deformation prediction model according to the root mean square error threshold, the mean absolute error threshold, and the determination coefficient threshold, and output the judgment result; S4. According to the judgment result, adjust the hyperparameters of the seismic sand liquefaction deformation prediction model or encapsulate the model, and retrain and update the seismic sand liquefaction deformation prediction model after adjusting the hyperparameters through the seismic sand liquefaction deformation data set; S5. Obtain new seismic sand liquefaction characteristic data, perform missing value processing, anomaly detection, feature selection processing, and standardization processing on the new seismic sand liquefaction characteristic data, and input it into the encapsulated seismic sand liquefaction deformation prediction model to obtain the seismic sand liquefaction deformation prediction result.

[0022] Further, the working principle of the present application is illustrated by the following embodiments: It is assumed that the present experiment focuses on a coastal city with frequent earthquakes, which has complex geological conditions, high groundwater level, and widespread distribution of sand.

[0023] Through the engineering investigation reports and monitoring data of historical seismic cases, 120 historical seismic cases in a seismic-prone area from 1980 to 2022 are collected, each case containing 20 sand liquefaction characteristic data and surface subsidence, each case is labeled with a unique ID and geographic location coordinates, the characteristic data and deformation are associated through the ID and geographic location coordinates to form a structured data set, and a multi-source coupled seismic data set is constructed.

[0024] The mean filling method is used for the multi-source coupled seismic data set. For example, when the clay content feature in the sand liquefaction characteristic data is missing, the mean value of the non-missing value of the clay content feature is calculated, assuming that the mean value is 15.2%, and the missing data is filled. The multi-source coupled seismic data set after the missing value processing is subjected to outlier detection. Taking the standard penetration number in the sand liquefaction characteristic data as an example, the IQR is calculated to be 4, the outlier range is 0 to 16, the cases with a standard penetration number less than 0 or greater than 16 are deleted, and a total of 5 abnormal samples are detected and discarded. The multi-source coupled seismic data set after the outlier detection is subjected to feature selection. The Spearman correlation coefficient of each feature and the ground settlement is calculated, the features with an absolute value of the correlation coefficient less than or equal to 0.1 are deleted, for example, the correlation coefficient of the soil particle specific gravity in the sand liquefaction characteristic data is 0.08, and it is deleted, and finally 14 key features are retained. The 14 key features are groundwater depth, soil layer depth, soil layer distribution, relative density, water content, effective overburden stress, clay content, standard penetration number, soil particle specific gravity, void ratio, peak ground acceleration, seismic duration, predominant period, velocity response spectrum value, magnitude and epicentral distance. The retained features are subjected to Z-score standardization. For example, the groundwater depth of a certain sample is 3.2 m, and the standardized value after standardization is -0.375.

[0025] The 120 samples are divided into a training set of 96 and a test set of 24 according to an 8:2 ratio. The XGBoost model is used, and the initial hyperparameters are set. In the hyperparameters, the number of trees is 100, the learning rate is 0.1, the maximum depth of the tree is 6, the sub-sampling ratio is 0.8, and the column sampling ratio is 0.8. The validation set loss is recorded during the training process. After 100 iterations, convergence is obtained, and a seismic sand liquefaction deformation prediction model is obtained. The test set is input into the trained seismic sand liquefaction deformation prediction model, the root mean square error RMSE of the seismic sand liquefaction deformation prediction model is calculated to be 2.15 cm, the mean absolute error MAE is 1.82 cm, and the determination coefficient R² is 0.72. The root mean square error RMSE, the mean absolute error MAE, and the determination coefficient R² of the seismic sand liquefaction deformation prediction model are compared with the respective thresholds. It is found that the root mean square error RMSE of the seismic sand liquefaction deformation prediction model is greater than 1, the mean absolute error MAE is greater than 0.8, and the determination coefficient R² is greater than 0.7. Since the RMSE and MAE do not meet the standard, the hyperparameter adjustment is triggered.

[0026] The hyperparameter search space is defined, in which the number of trees ranges from 50 to 200, the learning rate ranges from 0.01 to 0.2, the maximum depth of the tree ranges from 3 to 10, the subsampling ratio ranges from 0.6 to 1.0, and the column sampling ratio ranges from 0.6 to 1.0. After 20 iterations, the optimal hyperparameters are: the optimal number of trees is 150, the optimal learning rate is 0.05, the optimal maximum depth of the tree is 8, and the optimal subsampling ratio and column sampling ratio are both 0.7. The hyperparameters of the seismic sandy soil liquefaction deformation prediction model are adjusted according to the optimal parameters, and retrained. The root mean square error (MSE) of the retrained seismic sandy soil liquefaction deformation prediction model is 0.98 cm, the mean absolute error (MAE) is 0.75 cm, and the coefficient of determination (R²) is 0.85. The RMSE is less than 1, the MAE is less than 0.8, and the R² is greater than 0.7, which meets the deployment conditions.

[0027] The seismic sandy soil liquefaction deformation prediction model is packaged as a Web service interface through Flask. New site characteristic data is obtained, which has been processed for missing values, abnormal value detection, feature selection, and standardization. The site characteristic data is packaged into a data packet and sent to the server's application programming interface through an HTTP protocol POST request. The server obtains the data packet through the application programming interface and calls the seismic sandy soil liquefaction deformation prediction model packaged as a Web service interface. The seismic sandy soil liquefaction deformation prediction model outputs the seismic sandy soil liquefaction deformation prediction result according to the data packet, and returns a predicted ground settlement of 12.3 cm. The error is 0.9 cm after actual site monitoring verification, which meets the engineering application accuracy requirements.

[0028] It should be noted that the above-mentioned embodiments of the present application are only for description, not representing the advantages and disadvantages of the embodiments. The above describes specific embodiments of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or may be advantageous.

[0029] Each embodiment in the present specification is described in a progressive manner, and the same or similar parts of each embodiment can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0030] The above-described embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; modifying the technical solutions described in the above embodiments, or equivalently replacing some technical features, does not change the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A machine learning-based method for predicting earthquake-induced sand liquefaction deformation, characterized in that, Includes the following steps: S1. By using engineering survey reports and monitoring data of historical earthquake cases, obtain earthquake sand liquefaction characteristic data and sand liquefaction deformation amount of earthquake cases, and construct a multi-source coupled earthquake dataset accordingly; S2. The missing value processing of the multi-source coupled seismic dataset is performed. The multi-source coupled seismic dataset with missing value processing is then subjected to anomaly detection using IQR. Feature selection processing is then performed on the multi-source coupled seismic dataset with anomaly detection. At the same time, Z-score standardization is used to perform standardization processing to obtain the seismic sand liquefaction deformation dataset, which is then stored in the database. S3. Based on the earthquake sand liquefaction deformation dataset, construct an earthquake sand liquefaction deformation prediction model, calculate the root mean square error, mean absolute error, and coefficient of determination for the earthquake sand liquefaction deformation prediction model, judge the constructed earthquake sand liquefaction deformation prediction model according to the root mean square error threshold, mean absolute error threshold, and coefficient of determination threshold, and output the judgment result. S4. Based on the judgment results, the hyperparameters of the earthquake sand liquefaction deformation prediction model are adjusted or the model is encapsulated. The earthquake sand liquefaction deformation prediction model after hyperparameter adjustment is retrained and updated using the earthquake sand liquefaction deformation dataset. S5. Obtain new earthquake sand liquefaction characteristic data, perform missing value processing, outlier detection, feature selection processing, and standardization processing on the new earthquake sand liquefaction characteristic data, and input it into the encapsulated earthquake sand liquefaction deformation prediction model to obtain the earthquake sand liquefaction deformation prediction results.

2. According to claim 1, a method for predicting earthquake sand liquefaction deformation based on machine learning, characterized in that, The method for constructing a multi-source coupled seismic dataset: Based on engineering survey reports and testing data of historical earthquake cases, the following parameters were obtained: groundwater level depth, soil layer depth, soil layer distribution, relative density, water content, effective overburden stress, clay content, SPT blow count, soil particle specific gravity, void ratio, peak ground acceleration, earthquake duration, dominant period, velocity response spectrum, magnitude, epicentral distance, and surface subsidence. For each earthquake case, the geographical location information and unique ID number are manually labeled. Based on the geographical location information and unique ID number of the earthquake case, the earthquake sand liquefaction characteristic data and sand liquefaction deformation amount corresponding to the earthquake case are matched with the geographical location information and unique ID number of the earthquake case to establish the association and construct a multi-source coupled earthquake dataset. The earthquake-induced sand liquefaction characteristic data include groundwater level depth, soil layer burial depth, soil layer distribution, relative density, water content, effective overlying stress, clay content, SPT blow count, soil particle specific gravity, void ratio, peak ground acceleration, earthquake duration, dominant period, velocity response spectrum value, magnitude, and epicentral distance. The amount of sand liquefaction deformation is the amount of surface subsidence.

3. According to claim 2, a method for predicting earthquake sand liquefaction deformation based on machine learning, characterized in that, The method described above involves handling missing values ​​and detecting anomalies in a multi-source coupled seismic dataset, as well as standardizing the multi-source coupled seismic dataset after feature selection. Missing values ​​were filled using the mean imputation method for multi-source coupled seismic datasets; Outliers in the multi-source coupled seismic dataset after missing value processing are filtered using the IQR method and then discarded. Z-score normalization was performed on the source-coupled seismic dataset after feature selection.

4. According to claim 3, a method for predicting earthquake sand liquefaction deformation based on machine learning, characterized in that, The method described above involves feature selection processing of the multi-source coupled seismic dataset after anomaly detection and obtaining the seismic sand liquefaction deformation dataset: The Spearman correlation coefficient of all feature elements in the multi-source coupled seismic dataset after anomaly detection is calculated using the Spearman correlation coefficient formula. The correlation threshold is set to 0.

1. The absolute value of the Spearman correlation coefficient of all feature elements is compared with the correlation threshold. Feature elements with Spearman correlation coefficients less than or equal to the correlation threshold are deleted from the multi-source coupled seismic dataset. This is how feature selection processing of the multi-source coupled seismic dataset is achieved. The multi-source coupled seismic dataset that has undergone feature selection processing is standardized to obtain the seismic sand liquefaction deformation dataset. The characteristic elements are the groundwater level depth, soil layer depth, soil layer distribution, relative density, water content, effective overlying stress, clay content, SPT blow count, soil particle specific gravity, void ratio, peak ground acceleration, earthquake duration, dominant period, velocity response spectrum value, magnitude or epicentral distance of the earthquake sand liquefaction characteristic data.

5. A method for predicting earthquake sand liquefaction deformation based on machine learning, as described in claim 4, is characterized in that... The method for constructing an earthquake sand liquefaction deformation prediction model based on earthquake sand liquefaction deformation dataset: The earthquake sand liquefaction deformation dataset was obtained from the database. The dataset was divided into training and testing sets in an 8:2 ratio. An XGBoost model was built using the xgboost library, and the hyperparameters of the XGBoost model were set. The XGBoost model is trained using the training set to obtain a trained XGBoost model, which is a prediction model for earthquake sand liquefaction deformation. The xgboost library is a gradient boosting library; The XGBoost model is a machine learning algorithm. The hyperparameters include the number of trees, learning rate, maximum tree depth, subsampling ratio, and column sampling ratio.

6. According to claim 5, a method for predicting earthquake sand liquefaction deformation based on machine learning, characterized in that, The method described above involves judging the constructed earthquake sand liquefaction deformation prediction model based on the root mean square error threshold, the mean absolute error threshold, and the coefficient of determination threshold, and then outputting the judgment results. The test set is input into the earthquake sand liquefaction deformation prediction model. Based on the input data and the rules learned during the model training process, the model predicts the earthquake sand liquefaction deformation and outputs the prediction result, namely the predicted sand liquefaction deformation amount. Based on the predicted sand liquefaction deformation amount and the actual sand liquefaction deformation amount in the test set, the root mean square error, mean absolute error and coefficient of determination of the earthquake sand liquefaction deformation prediction model are obtained by using the root mean square error formula, the mean absolute error formula and the coefficient of determination formula, respectively. The root mean square error, mean absolute error, and coefficient of determination of the earthquake sand liquefaction deformation prediction model are compared with the root mean square error threshold, mean absolute error threshold, and coefficient of determination threshold, respectively. If the root mean square error of the earthquake sand liquefaction deformation prediction model is less than the root mean square error threshold, the mean absolute error is less than the mean absolute error threshold, and the coefficient of determination is greater than the coefficient of determination threshold, then the judgment result is to deploy the earthquake sand liquefaction deformation prediction model. Otherwise, the hyperparameters of the earthquake sand liquefaction deformation prediction model will be adjusted. The root mean square error threshold, the mean absolute error threshold, and the coefficient of determination threshold are 1, 0.8, and 0.7, respectively.

7. A method for predicting earthquake sand liquefaction deformation based on machine learning, as described in claim 6, is characterized in that... The method for adjusting the hyperparameters of the earthquake sand liquefaction deformation prediction model: The hyperparameters of the earthquake sand liquefaction deformation prediction model were adjusted using Bayesian optimization. The Bayesian optimization is a hyperparameter tuning method based on Bayes' theorem.

8. A method for predicting earthquake sand liquefaction deformation based on machine learning, as described in claim 7, is characterized in that... The method for encapsulating the earthquake sand liquefaction deformation prediction model: The earthquake sand liquefaction deformation prediction model is encapsulated into a web service interface using Flask, thereby realizing the model encapsulation of the earthquake sand liquefaction deformation prediction model. Flask is a lightweight Python web framework.

9. A method for predicting earthquake sand liquefaction deformation based on machine learning, as described in claim 8, is characterized in that... The method for obtaining the prediction results of earthquake-induced sand liquefaction deformation: The new seismic sand liquefaction feature data, which has undergone missing value processing, outlier detection, feature selection processing, and standardization processing, is packaged into a data packet in JSON format. The data packet is then sent to the seismic sand liquefaction deformation prediction model, which is encapsulated as a web service interface, via the HTTP protocol. At the same time, the seismic sand liquefaction deformation prediction model outputs the seismic sand liquefaction deformation prediction results and returns them via the HTTP protocol. The JSON format is a lightweight data exchange format; The HTTP protocol is a network protocol.

Images