Barrier dam stability evaluation method, system and equipment and storage medium

Abstract

The invention provides a barrier dam stability evaluation method, system and device and a storage medium, and belongs to the technical field of barrier dam stability evaluation.The barrier dam stability evaluation method includes the steps that a barrier dam data disaster database is established, stability influence factor analysis is conducted according to data in the barrier dam data disaster database, and a barrier dam stability evaluation result is obtained; obtaining characteristics meeting the stability of a plurality of barrage dams, and screening the data set according to the characteristics; performing class equilibrium synthesis on the screened barrier dam disaster database by using a synthetic minority class oversampling technology; training a random forest model by using data in the equalized barrier dam data set, and constructing a nonlinear relationship between the plurality of barrier dam stability characteristics and stability to obtain a barrier dam classification model; and acquiring parameters of a to-be-detected barrier dam and inputting the parameters into the barrier dam classification model to obtain a stability classification result of the to-be-detected barrier dam. By comprehensively considering influence factors, the accuracy of the model for evaluating the stability of the barrier dam is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of evaluation of the stability of barrier dams, and particularly relates to a method, a system, a device and a storage medium for evaluating the stability of barrier dams. Background Art

[0002] Barrier dams belong to natural dams, and they are very different from artificial earth-rock dams in terms of material composition and dam structure. They are generally formed by the rapid accumulation of landslide debris and soil particles, without going through the compaction process of artificial earth-rock dams. Therefore, there may be highly permeable areas composed of large boulder aggregates, and their impermeability and mechanical stability are poor. Due to their complex formation causes, the topographic and geomorphic features and the interaction between water and soil in the formation areas are different, resulting in a complex mechanism of action of the influencing factors on the stability of barrier dams, which has a significant impact on the stability of barrier dams. In addition, the stability of barrier dams is not only affected by the geomorphic and geological conditions of the area where they are located, but also related to the dam materials, structures and hydrodynamic conditions of the dams.

[0003] It is crucial to comprehensively consider the influencing factors of the stability of barrier dams for the rapid evaluation of the stability of barrier dams. Traditional methods quantitatively analyze the stability of barrier dams by establishing mathematical models, but most of them only establish linear models based on the influencing factors related to geomorphic features and ignore the complex non-linear relationship between the influencing factors of barrier dams and stability, resulting in the need to improve the accuracy of the models. Summary of the Invention

[0004] In order to solve the problems of inaccurate evaluation of the stability of barrier dams and incomplete consideration of influencing factors, the present invention provides a method, a system, a device and a storage medium for evaluating the stability of barrier dams.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] A method for evaluating the stability of a barrier dam, comprising the following steps:

[0007] Establish a barrier dam disaster database containing hydrodynamic characteristics, geomorphic characteristics and characteristic parameters of the barrier dam, and screen data that meet multiple stability characteristics of the barrier dam in the barrier dam disaster database;

[0008] Use the Synthetic Minority Over-sampling Technique (SMOTE) to perform class balance synthesis on the screened data to obtain a balanced barrier dam data set;

[0009] Use the data in the balanced barrier dam data set to train a random forest model, construct the non-linear relationship between the multiple stability characteristics of the barrier dam and the stability, and obtain a barrier dam classification model;

[0010] Obtain the characteristic parameters of the to-be-detected barrier lake dam and input them into the classification model of the barrier lake dam to obtain the stability classification result of the to-be-detected barrier lake dam, and evaluate the stability of the barrier lake dam according to the classification result.

[0011] Preferably, using the Synthetic Minority Over-sampling Technique (SMOTE) to perform class balance synthesis on the barrier lake dam disaster database, specifically, it is an over-sampling technique based on the K-Nearest Neighbor (KNN) algorithm, including the following steps:

[0012] (1) Set the number N of oversampled minority class samples;

[0013] (2) Calculate the distances of samples of the same type;

[0014] (3) Synthesize new samples, select the k nearest samples of the same type to the sample x i and randomly select samples to synthesize new samples according to the following formula:

[0015]

[0016] where rand(0, 1) represents generating a random number between (0, 1);

[0017] (4) Repeat the above processes (1)-(3) until the number of newly synthesized samples meets the balance requirement.

[0018] Preferably, in the random forest model, use Classification and Regression Tree (CART) to construct decision trees, calculate the Gini coefficient of the stability features through the CART, obtain the CART classification trees, and vote according to the classification results of each CART tree to obtain the classification result.

[0019] Preferably, the geomorphological features include the dam length, dam width, dam height, and dam body volume of the barrier lake dam; the hydrodynamic features include the storage capacity and basin area of the upstream barrier lake; the parameters of the barrier lake dam itself include the location, occurrence time, and stability of the barrier lake dam.

[0020] Preferably, the multiple stability features of the barrier lake dam specifically include dam height, dam length, dam width, dam body volume, storage capacity, basin area, material composition, and inducing factors.

[0021] Preferably, before performing class balance synthesis, it also includes encoding the inducing factors and material composition using one-hot encoding.

[0022] Preferably, before training the random forest model, it also includes parameter optimization of the number of trees n_estimators and the maximum depth max_depth of the trees in the forest of the random forest model through exhaustive search and 3-fold cross-validation.

[0023] The present invention also provides a stability evaluation system for a barrier dam, specifically including:

[0024] A data acquisition module, configured to establish a barrier dam disaster database including hydrodynamic characteristics, geomorphic characteristics, and characteristic parameters of the barrier dam, and screen data that meet multiple stability characteristics of the barrier dam in the barrier dam disaster database.

[0025] A data processing module, configured to use the synthetic minority over-sampling technique to perform class balance synthesis on the screened data to obtain a balanced barrier dam data set.

[0026] A model construction module, configured to use the data in the balanced barrier dam data set to train a random forest model, construct a non-linear relationship between the multiple stability characteristics of the barrier dam and the stability, and obtain a barrier dam classification model.

[0027] A stability evaluation module, configured to obtain the characteristic parameters of the barrier dam to be detected and input them into the barrier dam classification model, obtain the stability classification result of the barrier dam to be detected, and evaluate the stability of the barrier dam according to the classification result.

[0028] The present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory, and the processor executes the computer program to implement the steps in the above-mentioned method for evaluating the stability of a barrier dam.

[0029] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is loaded by a processor, it can execute the steps in the above-mentioned method for evaluating the stability of a barrier dam.

[0030] The method for evaluating the stability of a barrier dam provided by the present invention has the following beneficial effects:

[0031] By establishing a barrier dam disaster database and analyzing the influencing factors of the barrier dam stability, the present invention obtains multiple stability characteristics of the barrier dam, comprehensively considering the characteristics of the influencing factors of the stability. Using the synthetic minority over-sampling technique to perform class balance synthesis on the barrier dam data disaster database to obtain a balanced barrier dam data set, avoiding bias caused by data imbalance in the model, and strengthening the learning of the model for minority class samples. Training the random forest model to construct a non-linear relationship between multiple stability characteristics of the barrier dam and the stability, obtaining a barrier dam classification model, comprehensively considering the influencing factors of the barrier dam stability and the characteristics of the unbalanced distribution of the barrier dam categories, reflecting the complex non-linear relationship between the influencing factors of the barrier dam and the stability, and improving the accuracy of the evaluation of the barrier dam stability. Description of the Drawings

[0032] To more clearly illustrate the embodiments of the present invention and their design solutions, the accompanying drawings required for this embodiment will be briefly introduced below. The accompanying drawings in the following description are only partial embodiments of the present invention. For those of ordinary skill in the art, other accompanying drawings can be obtained based on these drawings without creative efforts.

[0033] Figure 1 It is a flowchart of a method for evaluating the stability of a barrier dam according to an embodiment of the present invention.

[0034] Figure 2 It is a schematic diagram of the process for evaluating the stability of a barrier dam in an embodiment of the present invention.

[0035] Figure 3 It is the relationship between the material composition and stability in an embodiment of the present invention.

[0036] Figure 4 It is the relationship between the inducing factors and the material composition in an embodiment of the present invention.

[0037] Figure 5 It is the relationship between the inducing factors and stability in an embodiment of the present invention.

[0038] Figure 6 It is the schematic diagram of oversampling principle of SMOTE algorithm in an embodiment of the present invention.

[0039] Figure 7 It is a flowchart of the random forest algorithm in an embodiment of the present invention.

[0040] Figure 8 It is a curve graph showing the change of average AUC with the number of decision trees in an embodiment of the present invention.

[0041] Figure 9 It is a curve graph showing the change of average AUC with the maximum depth of the tree in an embodiment of the present invention. Detailed implementation manners

[0042] To enable those skilled in the art to better understand the technical solutions of the present invention and implement them, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and cannot be used to limit the protection scope of the present invention.

[0043] Embodiment

[0044] The present invention provides a method for evaluating the stability of a barrier dam, as Figure 1 shown, which specifically includes the following steps:

[0045] S1. Construct a barrier lake disaster database. By collecting and statistically analyzing a series of databases with different characteristic parameters established by domestic and foreign scholars in the study of the stability and breach parameters of barrier lakes, a barrier lake disaster database containing 2,179 barrier lake cases was established. The establishment of the database mainly includes the following: basic information such as the country or region where the barrier lake is located, its name, occurrence time, and whether the barrier lake has formed and is stable. For geomorphological characteristics, the collected data includes the length, width, height, and volume of the barrier dam. For hydrodynamic characteristics, the storage capacity and catchment area of the upstream barrier lake are collected.

[0046] S2. Analyze the influencing factors of stability based on the data in the barrier lake disaster database to obtain multiple stability characteristics of barrier lakes. Select stable and unstable barrier lake cases with 8 characteristics including dam height, dam length, dam width, dam volume, storage capacity, catchment area, material composition, and inducing factors from the constructed barrier lake disaster database.

[0047] S3. Select barrier lake cases with all stability characteristic data from the established database as complete barrier lake research objects. A total of 96 barrier lake cases were selected, including 72 unstable cases and 24 stable cases. Some of the data are shown in Table 1.

[0048] Table 1 Specific data of some barrier lake cases

[0049]

[0050]

[0051] Among the 8 characteristics, 6 characteristics are continuous variables and 2 characteristics are discrete categorical variables. The stability label is a binary label. For categorical variables, random forests cannot directly use these characteristics for relevant calculations and need to encode the categorical variables. The one-hot encoding method is used to encode the inducing factors and material composition so that they are in the same position in the model calculation. At the same time, for the binary stability label, 1 represents the stable type and 0 represents the unstable type. After the above encoding operations, they can be imported into the model for calculation.

[0052] S4. Artificially synthesize the balanced screened barrier lake disaster database, use the synthetic minority over-sampling technique to balance and synthesize the data categories to obtain a balanced artificial virtual dataset, and select some data from the balanced artificial virtual dataset as the training set.

[0053] The synthetic minority over-sampling technique (SMOTE) is an over-sampling technique based on the K-nearest neighbor algorithm with strong anti-overfitting ability and is widely used in the problem of unbalanced sample classification. The basic idea is for each minority class sample xi , randomly select a sample from its nearest neighbors Then in x i and A point is randomly selected on the line between as a newly synthesized minority class sample, and the synthesized new sample is added to the original data set to achieve sample balance. Figure 6 This is a schematic diagram of the principle of SMOTE algorithm oversampling. The example sample has 2 features. The SMOTE algorithm process is as follows:

[0054] (1) Set the number of minority class samples after oversampling N. For example, there are 72 unstable dam samples and only 24 stable samples. The number of stable samples is relatively small. In order to obtain a balanced data set, the number of stable samples after oversampling is set to 72.

[0055] (2) Calculate the sample distance.

[0056] (3) Synthesize new samples. Select the same sample as x i The k closest samples of the same type are selected randomly from them Synthesize new samples according to the following formula:

[0057]

[0058] Among them, rand(0,1) means generating a random number between (0,1).

[0059] (4) Repeat the above process until the number of newly synthesized samples meets the requirement.

[0060] When there are categorical variables in the features, SMOTE supports specifying categorical features. Since categorical variables cannot be interpolated, SMOTE will refer to the features of the nearest neighbors of the new sample when synthesizing new samples, and then take the value with the highest number of occurrences.

[0061] In the original data set, there are 72 stable dam cases and 24 unstable cases, with a ratio of 3:1. It is an unbalanced data set with few stable cases. If the original data set is directly used to divide the training set and the test set, it will not only lead to insufficient utilization of minority class information during model training, reducing the evaluation accuracy of the model, but also too few stable cases in the divided test set will also reduce the reliability of the model. Therefore, the synthetic minority class oversampling technology (SMOTE) is used to balance the categories of the original data set to obtain a balanced artificial virtual data set containing 144 cases, including 72 unstable cases and 72 stable cases.

[0062] Considering that the number of samples in the dataset is small, if the dataset is divided in the ratio of 8:2, there are only 28 samples in the test set, and the reliability is insufficient. Therefore, in the present invention, according to the general division method, the dataset is randomly divided into a training set and a test set in the ratio of 7:3. During the process of selecting the feature for node splitting in the CART decision tree, the Gini coefficient of each feature is calculated separately, and then the feature with the smallest Gini coefficient is selected for splitting, without involving the mutual influence between features, and there is no weight problem caused by different orders of magnitude of feature vectors. Therefore, the data dimension does not affect the algorithm effect of the random forest, and the data used during training are all original data.

[0063] S5. Use the data in the training set to train the random forest RF model. As Figure 7 shown, use the basic learner CART decision tree of the random forest to construct multiple non-linear relationships between the stability characteristics of the barrier lake and the stability, as Figure 3 , Figure 4 and Figure 5 shown, and obtain the barrier lake classification model and evaluate it (SMOTE-RF algorithm model).

[0064] S51. CART (decision tree) is the basic component unit of the random forest (RF). In the classification problem, the recursive construction CART classification algorithm is used for the dataset. The algorithm starts from the root node. For the dataset D, if the amount of data in D is less than the sample number threshold or the sample lacks features, the subtree is returned and the recursion stops at the current node position. When the CART tree performs feature splitting, it mainly judges according to the Gini coefficient. The larger the Gini coefficient, the higher the uncertainty of the feature, and the smaller the Gini coefficient, the lower the uncertainty of the feature. Suppose there are M categories in total, and the probability of the m-th category is P m , and the calculation formula of the Gini coefficient is shown in Equation (2).

[0065]

[0066] For the data set S in the data set D m If the output number of the m-th category is V' and the probability of this category in D is V D , then the Gini coefficient of the data set D can be expressed as Equation (3).

[0067]

[0068] For the feature A in the data set D, the data set D is divided into D 1 and D 2 two data sets. Then, in the case of feature A, the Gini coefficient of the data set D can be calculated by Equation (4).

[0069]

[0070] In the formula, is the quantity of feature A in dataset D 1 ; is the quantity of feature A in dataset D2. Substituting formula (3) into formula (4) can obtain formula (5).

[0071]

[0072] In the formula, is the quantity of feature A in the m-th class of dataset D 1 ; is the quantity of feature A in the m-th class of dataset D2.

[0073] According to the Gini coefficient calculation formula, calculate the Gini coefficient of the data subset. If the Gini coefficient is less than the set threshold, stop recursion and return the CART tree subtree; at the same time, calculate the Gini coefficient of each feature value in the current node, and select the feature and its value with the lowest uncertainty degree in the Gini coefficient as the optimal splitting point and feature; split the dataset with the optimal splitting point and feature to obtain 2 data subsets, and generate the corresponding child nodes; repeat the above steps for the generated child nodes to obtain the final CART classification tree.

[0074] S52. The present invention establishes a random forest evaluation model based on the python platform. Before establishing the model, it is necessary to optimize the parameters of the random forest algorithm. Among the numerous hyperparameters of the random forest, 2 relatively important parameters are the number of trees n_estimators in the forest and the maximum depth max_depth of the tree. The random forest is an ensemble model composed of numerous decision trees. The number of trees in the forest greatly affects the performance of the model. On the one hand, if the number of trees in the forest is too small, the model training degree is insufficient, the information utilization rate of the training data is insufficient, which may lead to the model being unable to converge and the model being in an "underfitting" state; on the other hand, if the number of decision trees is too large, it will increase the complexity of the model, but cannot improve the performance of the model, resulting in a lower efficiency of the model. The maximum depth of the tree can be considered as the complexity of the model. If the depth is small, the model is relatively simple and the fitting effect is very poor. If the depth is too large, the model is complex and prone to overfitting.

[0075] In order to accurately determine the hyperparameters, use the method of traversing optimization combined with 3-fold cross-validation to determine the optimal hyperparameters of the random forest. The specific parameter definitions and adjustment ranges are shown in Table 2.

[0076] Table 2 Random forest parameter definitions and adjustment ranges

[0077]

[0078] Such as Figure 8The result graph for searching the optimal n_estimators using the traversal search method, that is, the corresponding relationship between the average AUC of 3-fold cross-validation on the training set and the number of decision trees in the forest. It can be seen from the graph that when the number of decision trees is small, the average AUC of cross-validation is significantly low, indicating that the model is not fully trained. When the number of decision trees increases to a certain extent, the average AUC fluctuates within a small range but is basically stable. The optimal number of decision trees for this optimization is 271, and the average AUC of cross-validation is 0.881.

[0079] (2) Search for the optimal max_depth.

[0080] Set the parameter in the random forest to the optimal n_estimators = 271 after optimization, and search for the optimal max_depth. The search results are as Figure 9 shown. When max_depth is small, the average AUC is significantly low, indicating that the model complexity is low at this time and the model is in an "underfitting" state. As this value increases, the model fitting degree gets better and then tends to be stable. The optimal max_depth = 9 is obtained from the curve, and the average AUC is 0.881 at this time. Within the selected range, the random forest model does not show overfitting, but it also achieves the optimal model performance.

[0081] The optimal parameters determined by traversal search and 3-fold cross-validation are shown in Table 3, and the other parameters are default values. Set the random forest parameters to the optimal parameters and retrain the model on the training set to obtain the optimal SMOTE-RF model.

[0082] Table 3 Optimal parameters of the SMOTE-RF model

[0083]

[0084] S6. In a binary classification problem such as judging the stability of a barrier dam (stable or unstable), the output result of the model needs to be compared with the actual situation to evaluate various performance aspects of the model prediction, such as accuracy and reliability.

[0085] Binary classification is a problem with only 2 classes in the classification task, and the output result of the model has only two values (0 and 1). For a specific binary classification problem, evaluating the performance of the model is a key step in solving the problem. Combine the true class and the model evaluation class as:

[0086] True positive (TP): The true class is positive and the evaluation class is positive; True negative (TN): The true class is negative and the evaluation class is negative; False positive (FP): The true class is negative and the evaluation class is positive; False negative (FN): The true class is positive and the evaluation class is positive.

[0087] The matrix composed of the above four cases is called the confusion matrix, as shown in Table 4.

[0088] Table 4 Confusion Matrix

[0089]

[0090] The confusion matrix is the most intuitive method to evaluate the performance of a classification model. In addition, there are several other indicators to measure the model performance:

[0091] (1) Accuracy: Accuracy is an indicator that measures the model performance from a global perspective and reflects the overall prediction correctness of the model. As shown in Equation (6).

[0092]

[0093] (2) Precision: Among the samples evaluated as positive examples, the proportion of actual positive examples, or among the samples evaluated as negative examples, the proportion of actual negative examples, which reflects the prediction accuracy of stable barrier lakes, as shown in Equation (7).

[0094]

[0095] (3) Recall: Among the samples that are actually positive examples, the proportion of correct evaluations, or among the samples that are actually negative examples, the proportion of correct evaluations, which can reflect the coverage of the model for unstable barrier lakes, as shown in Equation (8).

[0096]

[0097] (4) F1-Score: It comprehensively considers precision and recall, and when both reach a relatively high level, it takes their balanced value, as shown in Equation (9).

[0098]

[0099] (5) ROC (Receiver Operating Characteristic) Curve: For a binary classification model, based on different classification thresholds, different classification results are obtained, and for each classification result, two indicators, the True Positive Rate (TPR) and the False Positive Rate (FPR), can be calculated. TPR represents the proportion of truly positive example samples that are correctly classified, as shown in Equation (10).

[0100]

[0101] The False Positive Rate (FPR) represents the proportion of negative samples misclassified as positive by the classifier, as shown in Equation (11).

[0102]

[0103] Using these different classification results, an ROC curve is constructed. The curve takes the FPR as the abscissa and the TPR as the ordinate, and a series of coordinate points are obtained through these different classification results. The evaluation of the ROC curve has strong stability. The closer the curve is to the upper left corner, the greater the coverage rate of the true positive samples in the classification results, the fewer negative samples misjudged as positive, and the stronger the performance of the model.

[0104] (6) AUC (Area Under Curve): The value of AUC represents the area under the ROC curve and is often used to evaluate the performance of classification models. The closer the value of AUC is to 1, the better. Generally speaking, when AUC is close to 0.5, it means that the evaluation effect of this classification model is no different from random guessing, and when the AUC value is less than 0.5, it indicates that the effect of this classification model is worse than random guessing. In the evaluation of the stability of barrier dams, the ROC curve and AUC can help select appropriate models and compare the advantages and disadvantages of different models.

[0105] The SMOTE-RF model was applied to the stability evaluation of 96 real barrier dam cases with complete data, and the actual observed stability was used as a reference. The confusion matrix constructed by the model evaluation stability and the actual observed stability is shown in Table 5.

[0106] Among the 72 unstable barrier dam cases, only 1 case was misjudged as stable, and among the 24 stable barrier dam cases, only 3 cases were misjudged as unstable. This shows that the SMOTE-RF model constructed using the synthetic virtual dataset can relatively accurately evaluate the stability of 96 barrier dam cases, with an overall accuracy rate of 95.8%. The model evaluation results show good consistency with the actual observed stability.

[0107] Table 5 Stability evaluation results of 96 real barrier dam cases

[0108]

[0109] Compare the evaluation results of the model established using all feature parameters with the model established only using the geomorphological features of the barrier dam. The comparison results are shown in Table 6.

[0110] As shown in Table 6, the overall evaluation accuracy rate of the model established only using geomorphological features is 85.4%, the evaluation accuracy rate for stable cases is 70.8%, and the evaluation accuracy rate for unstable cases is 90.3%. Compared with the model established using all features, all aspects of indicators have decreased to a certain extent, considering that some influencing factors cannot fully reflect the information related to the stability of barrier lakes contained in the data.

[0111] Table 6 Comparison of the evaluation results of the models established using all features and geomorphological features

[0112]

[0113] Comparison of the results of applying the traditional barrier lake stability evaluation method and the model of the present invention to evaluate the stability of 96 barrier lake cases. As shown in Table 7, the overall accuracy rate of the SMOTE-RF model reaches 95.8%, while the overall accuracy rate of the traditional evaluation method is between 6.5% and 75%.

[0114] Table 7 Comparison between the SMOTE-RF model and the traditional barrier lake stability evaluation method

[0115]

[0116]

[0117] The calculation results of each evaluation method are compared and analyzed using the misjudgment rate F, the conservative accuracy rate Rc, and the absolute accuracy rate R. The results are shown in Table 8. The misjudgment rate of the SMOTE-RF model is only 1.0%, the conservative accuracy rate is 99.0%, and the absolute accuracy rate is 95.8%. Therefore, using the SMOTE-RF model to evaluate the stability of barrier lakes is more accurate.

[0118] Table 8 Comparison of the results between the SMOTE-RF model and each evaluation method

[0119]

[0120] S7. Obtain the characteristic parameters of the barrier lake to be detected and input them into the barrier lake classification model to obtain the stability classification result of the barrier lake to be detected.

[0121] The present invention also provides a barrier lake stability evaluation system, specifically including:

[0122] A data acquisition module, used to establish a barrier lake disaster database containing hydrodynamic characteristics, geomorphological characteristics, and barrier lake characteristic parameters, and screen data that meet multiple barrier lake stability characteristics in the barrier lake disaster database.

[0123] A data processing module, which is used to perform class - balanced synthesis on the screened data by using the Synthetic Minority Over - sampling Technique (SMOTE) to obtain a balanced barrier lake dataset.

[0124] A model construction module, which is used to train a random forest model with the data in the balanced barrier lake dataset, construct the non - linear relationship between multiple barrier lake stability characteristics and stability, and obtain a barrier lake classification model.

[0125] A stability evaluation module, which is used to obtain the characteristic parameters of the barrier lake to be detected and input them into the barrier lake classification model, obtain the stability classification result of the barrier lake to be detected, and evaluate the stability of the barrier lake according to the classification result.

[0126] Each module in the above - mentioned barrier lake stability evaluation system can be implemented in whole or in part by software, hardware, or a combination thereof. The above - mentioned modules can be embedded in the processor of a computer device in hardware form or be independent of it, or can be stored in the memory of a computer device in software form, so as to facilitate the processor to call and execute the operations corresponding to the above - mentioned modules.

[0127] The present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the steps in an embodiment of a method for evaluating the stability of a barrier lake. The specific implementation method can refer to the method embodiment and will not be elaborated here.

[0128] Furthermore, the present invention also provides a non - transitory computer - readable storage medium containing instructions, on which a computer program is stored. For example, a memory containing instructions. The above - mentioned instructions can be executed by the processor of a computer device to complete the above - mentioned method. For example, the non - transitory computer - readable storage medium can be a ROM, a random access memory (RAM), a CD - ROM, magnetic tape, floppy disk, and optical data storage device, etc. When the computer program is executed by the processor, it can implement the steps in an embodiment of a method for evaluating the stability of a barrier lake. The specific implementation method can refer to the method embodiment and will not be elaborated here.

[0129] Those skilled in the art should understand that the embodiments of the present invention can provide a method, a system, or a computer program product. Therefore, the present invention can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can be in the form of a computer program product implemented on one or more computer - usable storage media (including but not limited to disk storage, CD - ROM, optical storage, etc.) containing computer - usable program code.

[0130] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0131] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0132] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0133] It should be pointed out that the specific implementation methods described above can enable those skilled in the art to understand the invention more comprehensively, but do not limit the invention in any way. Therefore, although the invention has been described in detail in this specification and embodiments, those skilled in the art should understand that the invention can still be modified or replaced by equivalents; and all technical solutions and improvements that do not deviate from the spirit and scope of the invention are included in the protection scope of the patent for the invention. Any figure mark in the claims should not be regarded as limiting the claims involved. Any simple change or equivalent replacement of the technical solution that can be obviously obtained by any technician familiar with the field within the technical scope disclosed in the present invention belongs to the protection scope of the present invention.

Claims

1. A method for evaluating the stability of a landslide dam, characterized in that: The following steps are involved: Establishing a barrier dam disaster database including hydrodynamic characteristics, geomorphological characteristics and barrier dam characteristic parameters, and screening data satisfying multiple barrier dam stability characteristics in the barrier dam disaster database; The synthetic minority class oversampling technique is used to perform balanced synthesis on the filtered data to obtain a balanced dam dataset. Using the data in the balanced dam dataset to train the random forest model, constructing a nonlinear relationship between the stability characteristics and stability of the plurality of dams, and obtaining a dam classification model; The characteristic parameters of the dam to be detected are obtained and input into the dam classification model to obtain the stability classification result of the dam to be detected, and the stability of the dam is evaluated according to the classification result.

2. A method for evaluating the stability of a landslide dam according to claim 1, characterized in that: The method of using the synthetic minority class oversampling technology to perform class balanced synthesis on the barrier dam disaster database is specifically an oversampling technology based on the K nearest neighbor algorithm, and includes the following steps: (1) Set the number N of minority class samples after oversampling; (2) Calculate the distance between samples of the same type; (3) Synthesize new samples and select the same samples as sample x i The k closest samples of the same type, randomly selected samples Synthesize new samples according to the following formula: Among them, rand(0,1) means generating a random number between (0,1); (4) Repeat the above process (1)-(3) until the number of newly synthesized samples meets the balance requirement.

3. A method for evaluating the stability of a landslide dam according to claim 1, characterized in that: In the random forest model, a classification and regression tree CART is used to construct a decision tree, and the Gini coefficient of the stability feature is calculated by the classification and regression tree CART to obtain a CART classification tree. Voting is performed according to the classification results of each CART tree to obtain a classification result.

4. A method for evaluating the stability of a landslide dam according to claim 1, characterized in that: The geomorphological characteristics include the length, width, height and volume of the dam; the hydrodynamic characteristics include the storage capacity and basin area of ​​the upstream dam lake; the parameters of the dam itself include the area where the dam is located, the time of occurrence and whether the dam is stable.

5. A method for evaluating the stability of a landslide dam according to claim 1, characterized in that: The stability characteristics of the multiple landslide dams specifically include dam height, dam length, dam width, dam volume, reservoir capacity, basin area, material composition and inducing factors.

6. A method for evaluating the stability of a landslide dam according to claim 5, characterized in that: It also includes encoding the inducing factors and material composition using one-hot encoding before performing category-balanced synthesis.

7. A method for evaluating the stability of a landslide dam according to claim 1, characterized in that: Before training the random forest model, the method also includes optimizing the number of trees n_estimators and the maximum depth max_depth of the trees in the forest in the random forest model by traversal search and 3-fold crossover.

8. A dam stability evaluation system, characterized in that: include: A data acquisition module is used to establish a barrier dam disaster database including hydrodynamic characteristics, geomorphological characteristics and barrier dam characteristic parameters, and to screen data satisfying multiple barrier dam stability characteristics in the barrier dam disaster database; A data processing module is used to perform balanced synthesis of the filtered data using synthetic minority class oversampling technology to obtain a balanced dam dataset; A model building module is used to train a random forest model using data in the balanced dam dataset, build a nonlinear relationship between the stability characteristics and stability of the plurality of dams, and obtain a dam classification model; The stability evaluation module is used to obtain characteristic parameters of the dam to be detected and input them into the dam classification model to obtain the stability classification result of the dam to be detected, and evaluate the stability of the dam according to the classification result.

9. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is loaded into a processor, the computer program can execute the steps of the method according to any one of claims 1 to 7.

Images