A method and system for rapid assessment of slope stability

Abstract

This invention belongs to the field of slope stability analysis technology, and provides a method and system for rapid slope stability assessment. It involves collecting slope engineering cases through field exploration or literature review, extracting feature information, and establishing a case dataset. Statistical analysis is performed on the parameters in the case dataset, and corresponding numerical simulation schemes are designed to obtain numerical simulation results. A numerical simulation database is constructed to enhance the data. A slope stability evaluation model is built based on a classification neural network algorithm. The case dataset is divided into training and testing sets, and the numerical simulation database is integrated into the training set to construct an enhanced training set. The slope stability evaluation model is trained based on the enhanced training set, and its performance is evaluated using the testing set. Feature information of the target slope is obtained, and the optimal slope stability evaluation model after evaluation is used to process the information, yielding the slope stability assessment result. This invention enables rapid prediction of slope stability.

Description

Technical Field

[0001] This invention belongs to the field of slope stability analysis technology, specifically relating to a rapid slope stability assessment method and system. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Slope stability analysis is one of the core research directions in geotechnical engineering. In various engineering practices such as mining, highway and railway construction, water conservancy projects, and building construction, numerous artificial or natural slopes are formed, and their stability directly affects project safety. Stability analysis, as a core component of slope design and research, is of decisive significance for project safety. Once slope instability occurs, it can not only lead to significant economic losses such as project shutdowns and equipment damage, but also pose a serious threat to people's lives. Therefore, constructing slope stability prediction models that are both reliable and effective is of significant practical value for ensuring project safety and reducing disaster risks.

[0004] Traditional slope stability analysis methods mainly include limit equilibrium method, finite element method, and finite difference method, but all of them have obvious limitations: the limit equilibrium method has to deal with the problem of identifying a large number of potential sliding surfaces in the slope, and the process of finding the critical sliding surface is extremely complex; numerical simulation method is highly dependent on the setting of boundary conditions and the selection of mechanical parameters, which not only requires researchers to have rich engineering experience and field analysis capabilities to ensure the rationality of the results, but also the simulation process is time-consuming and difficult to meet the needs of rapid evaluation. Summary of the Invention

[0005] To address the aforementioned problems, this invention proposes a rapid slope stability assessment method and system. This invention enables rapid prediction of slope stability and improves the reliability of slope stability assessment.

[0006] According to some embodiments, the present invention adopts the following technical solution: A rapid slope stability assessment method includes the following steps: Collect slope engineering cases through field exploration or literature review, extract characteristic information, and establish a case dataset; Statistical analysis is performed on the parameters in the case dataset, corresponding numerical simulation schemes are designed, numerical simulation results are obtained, and a numerical simulation database is constructed to achieve data augmentation. A slope stability evaluation model is constructed based on a classification neural network algorithm. The case dataset is divided into a training set and a test set. A numerical simulation database is integrated into the training set to construct an enhanced training set. The slope stability evaluation model is trained based on the enhanced training set, and the performance of the slope stability evaluation model is evaluated using the test set. The characteristic information of the target slope is obtained, and the optimal slope stability evaluation model after evaluation is used to process it to obtain the evaluation result of slope stability.

[0007] As an alternative implementation, the feature information includes the unit weight of the slope material, cohesion, internal friction angle, slope height, slope angle, and pore pressure ratio; wherein, the pore pressure ratio is defined as the ratio of the pore water pressure at a certain point in the slope to its vertical self-weight stress.

[0008] As an alternative implementation method, the process of designing the corresponding numerical simulation scheme includes: on the premise that the mean and standard deviation of each feature parameter in the numerical simulation database and the case dataset are consistent, and the maximum and minimum values ​​are similar, a normal distribution sampling method is used to generate the numerical simulation scheme.

[0009] As a further defined implementation method, the numerical simulation results are obtained through batch calculations.

[0010] As an alternative implementation method, the process of constructing a slope stability evaluation model includes: utilizing a probabilistic neural network, which comprises an input layer, a pattern layer, a summation layer, and an output layer. The input layer is responsible for reading learning samples from a large database of slope stability data; the pattern layer is trained on the data features of the learning samples to extract the patterns and relationships contained in the samples, providing a basis for the identification and classification of slope stability; the summation layer performs summation operations on the pattern vectors of each category to complete the probability density estimation of each category; and the output layer determines the final classification result based on the category with the highest probability density. A set of intelligent identification models for slope stability with different combinations of input parameters is generated using probabilistic neural networks.

[0011] As a further defined implementation, each neural network unit in the summing layer is connected only to the corresponding type of model neuron and is estimated according to the Parzen window function method and various types of conditional probability densities.

[0012] As an alternative implementation, the process of training the slope stability evaluation model based on the enhanced training set includes training each model generated using a probabilistic neural network using the enhanced training set.

[0013] As an alternative implementation method, the process of using a test set to complete the performance evaluation of the slope stability evaluation model includes selecting the model with the best prediction accuracy. When considering the characteristics of unit weight, cohesion, internal friction angle, slope angle, slope height and pore pressure ratio, the model with the highest prediction accuracy is selected. The performance of this model is then evaluated, and the evaluation parameters include accuracy, precision, recall and F1-score.

[0014] A rapid slope stability assessment system includes: The feature information extraction module is configured to collect slope engineering cases through field exploration or literature review, extract feature information, and establish a case dataset. The data augmentation module is configured to perform statistical analysis on the parameters in the case dataset, design corresponding numerical simulation schemes, obtain numerical simulation results, and build a numerical simulation database to achieve data augmentation. The model building and training module is configured to build a slope stability evaluation model based on a classification neural network algorithm. The case dataset is divided into a training set and a test set. A numerical simulation database is integrated into the training set to build an enhanced training set. The slope stability evaluation model is trained based on the enhanced training set, and the performance of the slope stability evaluation model is evaluated using the test set. The stability assessment module is configured to acquire the characteristic information of the target slope, process it using the optimal slope stability evaluation model after assessment, and obtain the slope stability assessment result.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention constructs an intelligent slope stability identification model based on a classification neural network algorithm, which has strong interpretability and predictive stability. The model's computational performance on different datasets is stable, and the model parameters are highly interpretable, enabling in-depth understanding and learning of the slope stability identification process.

[0016] This invention employs multiple models to perform calculations and predict results on a dataset, and then calculates the prediction accuracy of each model. Based on a probabilistic neural network, it conducts a comprehensive analysis of all possible combinations of input parameters. Compared to other prediction models, this invention is more comprehensive in terms of parameter combination coverage and analytical depth, enabling it to capture data features more fully and thus improve the accuracy and reliability of predictions.

[0017] This invention provides a rapid intelligent evaluation modeling method for slope stability that integrates case datasets and numerical simulation enhancement. Compared with traditional analysis methods, this invention innovatively introduces a numerical simulation database into the training set of the neural network, thereby significantly improving the prediction accuracy of the model. This method improves the accuracy of the model's prediction results.

[0018] This invention proposes a method for designing numerical simulation schemes. Under the premise of ensuring that the mean and standard deviation of each feature parameter in the numerical simulation database and the case dataset are consistent, and that the maximum and minimum values ​​are similar, the method uses a normal distribution to generate numerical simulation schemes, thereby improving the accuracy and efficiency of data augmentation.

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0020] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0021] Figure 1 This is the method flow of an embodiment of the present invention; Figure 2 This is a schematic diagram of the case dataset in an embodiment of the present invention; Figure 3 The image shown is a violin diagram of the case dataset in this embodiment of the invention. Figure 4 This is a correlation matrix diagram of the case dataset in this embodiment of the invention; Figure 5 This is a correlation pairing graph of the case dataset in this embodiment of the invention; Figure 6 This is a schematic diagram of the numerical simulation database in an embodiment of the present invention; Figure 7 This is a schematic diagram of the parameters corresponding to the intelligent identification model in an embodiment of the present invention; Figure 8 This is a diagram showing the test results of the probabilistic neural network in an embodiment of the present invention; Figure 9 This is the optimal model evaluation diagram in the embodiments of the present invention. Detailed Implementation

[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0023] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0024] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0025] Where there is no conflict, the embodiments and features described in this application may be combined with each other.

[0026] Example 1 A rapid slope stability assessment method combining engineering case studies and numerical simulation, such as Figure 1 The above includes the following steps: Step 1: Collect slope engineering cases based on field exploration or literature research and obtain relevant engineering characteristic information such as unit weight of slope materials, cohesion, internal friction angle, and slope angle to establish a case dataset.

[0027] In this example, the necessary information to be collected includes: (1) unit weight; (2) cohesion; (3) internal friction angle; (4) slope angle; (5) slope height; (6) pore pressure ratio; and (7) slope stability.

[0028] To verify the technical effect of the present invention, based on 684 slope cases in the case dataset, each data group contains 7 engineering characteristic information of the slope: unit weight ( ), cohesion ( ), internal friction angle ( ), slope angle ( ), slope height ( ), pore pressure ratio ( ), slope stability. Detailed information for the selected case dataset is as follows: Figure 2 As shown.

[0029] Step 2: Perform statistical analysis on the parameters in the case dataset, design corresponding numerical simulation schemes. Given the large number of simulation schemes, batch calculation is used to obtain numerical simulation results, thereby forming a numerical simulation database.

[0030] In this embodiment, the data distribution of each parameter in the dataset is visualized in the form of a violin plot (e.g., Figure 2 As shown), the correlation of each parameter is presented in a correlation matrix and a correlation pairing plot (e.g., Figure 3 , Figure 4(As shown). Violin plots provide richer information about data distribution, facilitating intuitive and clear analysis of data characteristics, such as identifying multimodal distributions or outliers. In a violin plot, the bottom and top represent the first and third quartiles of the data, respectively. Furthermore, the width of the violin plot reflects the density distribution of the data; a wider plot indicates higher data density in that region. Correlation matrix plots present the correlation between various parameters, revealing weak correlations. Correlation pair plots visualize the correlation between different groups of data; the kernel density plot along the main diagonal reveals the distribution characteristics of each data point. Based on the analysis of violin plots and correlation pair plots, combined with the database's maximum, minimum, mean, and standard deviation, a numerical simulation scheme was determined, comprising 200 groups. Given the large number of numerical simulations, batch calculations were performed to obtain the simulation results and effectively reduce simulation time. The numerical simulation scheme and results are attached. Figure 5 As shown.

[0031] Step 3: Construct a slope stability evaluation model based on a classification neural network algorithm. The case dataset is divided into a training set (80%) and a test set (20%). An enhanced training set is constructed by incorporating a numerical simulation database into the original training set. The neural network is trained based on the enhanced training set, and the model performance is evaluated using the test set. Finally, a reliable slope stability evaluation model is established.

[0032] In this embodiment, the parameter input order of the slope stability intelligent identification model learning samples is: unit weight ( ), cohesion ( ), internal friction angle ( ), slope angle ( ), slope height ( ), pore pressure ratio ( The pore pressure ratio is defined as the ratio of the pore water pressure at a point on the slope to the vertical self-weight stress at that point (calculated in a semi-infinite space). In engineering, for ease of calculation, the pore pressure ratio is usually determined by the ratio of the buoyancy force on the entire sliding body to its total weight. In this embodiment, the pore pressure ratio in the numerical simulation scheme is obtained using the total stress method. The output parameters of the intelligent slope stability identification model are slope stability... sex.

[0033] The intelligent slope stability identification model in this invention utilizes a probabilistic neural network, which is the core of the intelligent slope stability identification model.

[0034] A probabilistic neural network consists of four layers: an input layer, a pattern layer, a summation layer, and an output layer.

[0035] ① Input layer: The input layer has 7 neurons, each corresponding to a different sample. The neurons directly transmit the information received from the input layer to the pattern layer.

[0036] ② Pattern layer: The pattern layer has 660 neurons, and each neuron in the pattern layer corresponds to a different learning sample. The transfer function formula (1) for the neurons in the pattern layer is as follows: (1) in, Indicates network input variables; For the first The learning samples corresponding to each neuron These are the function control parameters.

[0037] ③ Summation Layer: Each neural network unit is connected only to the corresponding type of model neuron, and the probability density is estimated according to the Parzen window function method and various types of conditional probability densities. The Parzen window probability density estimation is as follows: (2) in, This indicates the number of input nodes in the neural network. This represents the number of training samples in the input layer of the neural network. This represents the hyperparameters of a neural network. This represents the input vector of the neural network. Indicates the neural network's first... The input vector of each training sample.

[0038] ④ Output layer: The output layer has 1 neuron, and each neuron represents a category. The most likely category is selected as the output result through a competition mechanism.

[0039] To compare the influence of different parameters on the intelligent identification model of slope failure and instability modes, a probabilistic neural network was used to generate a set of intelligent identification models for slope stability with different combinations of input parameters. The corresponding input parameters for each model are shown below. Figure 6 As shown.

[0040] Based on the 63 models generated by the neural network, each model was first trained using the original training set, and then trained using the augmented training set. The test results are shown in the figure. Figure 7 .Depend on Figure 7It can be observed that the overall prediction performance of the original training set is poor. In terms of prediction performance, the model using only two input parameters has the highest accuracy, and the model with fewer input parameters has even higher prediction accuracy. As the number of input parameters increases, the average prediction accuracy of the probabilistic neural network first increases and then decreases. Conversely, for the augmented training set, when there are fewer than four input parameters, its prediction performance is not significantly different from the original training set; however, when there are more than four input parameters, the prediction performance of the augmented training set is significantly improved. The model on this training set achieves the highest prediction accuracy (88.3%) when using all input parameters, and the average prediction accuracy increases with the number of input parameters.

[0041] The experimental results above show that adding the numerical simulation database to the neural network training set and retesting the model can significantly improve the prediction accuracy.

[0042] Step 4: Comparative analysis revealed that introducing a numerical simulation database into the training set significantly improved the model's prediction accuracy. Comparative results based on relevant evaluation parameters demonstrate that the model possesses good applicability and reliability.

[0043] The characteristic information of the target slope is obtained, and the optimal slope stability evaluation model after evaluation is used to process it to obtain the evaluation result of slope stability.

[0044] In this embodiment, the model with the best prediction accuracy is selected, i.e., the model with the highest prediction accuracy when considering six parameters: unit weight, cohesion, internal friction angle, slope angle, slope height, and pore pressure ratio. The model's performance is then evaluated, with parameters including accuracy, precision, recall, and F1-score. The corresponding confusion matrix and recall and precision are shown below. Figure 8 As shown.

[0045] Figure 8 The selected model is the 63rd model of the probabilistic neural network, which considers unit weight ( ), cohesion ( ), internal friction angle ( ), slope angle ( ), slope height ( ), pore pressure ratio ( The model with these six input parameters is composed of... Figure 9 It can be observed that the model's precision and recall are both above 80% for both unstable and stable cases, indicating that the model's overall ability to classify slope stability is good. Furthermore, the model's F1 score of 0.88 further demonstrates its high applicability.

[0046] Example 2 A rapid slope stability assessment system includes: The feature information extraction module is configured to collect slope engineering cases through field exploration or literature review, extract feature information, and establish a case dataset. The data augmentation module is configured to perform statistical analysis on the parameters in the case dataset, design corresponding numerical simulation schemes, obtain numerical simulation results, and build a numerical simulation database to achieve data augmentation. The model building and training module is configured to build a slope stability evaluation model based on a classification neural network algorithm. The case dataset is divided into a training set and a test set. A numerical simulation database is integrated into the training set to build an enhanced training set. The slope stability evaluation model is trained based on the enhanced training set, and the performance of the slope stability evaluation model is evaluated using the test set. The stability assessment module is configured to acquire the characteristic information of the target slope, process it using the optimal slope stability evaluation model after assessment, and obtain the slope stability assessment result.

[0047] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of one or more computer-usable storage media (including, but not limited to, disk storage, etc.) containing computer-usable program code. CD - ROM It takes the form of a computer program product implemented on (such as optical memory, etc.).

[0048] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0049] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The function specified in one or more boxes.

[0050] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0051] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art without creative effort within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A rapid assessment method for slope stability, characterized in that, Includes the following steps: Slope engineering cases were collected through field exploration or literature review, feature information was extracted, and a case dataset was established. Statistical analysis is performed on the parameters in the case dataset, corresponding numerical simulation schemes are designed, numerical simulation results are obtained, and a numerical simulation database is constructed to achieve data augmentation. A slope stability evaluation model is constructed based on a classification neural network algorithm. The case dataset is divided into a training set and a test set. A numerical simulation database is integrated into the training set to construct an enhanced training set. The slope stability evaluation model is trained based on the enhanced training set, and the performance evaluation of the slope stability evaluation model is completed using the test set. The characteristic information of the target slope is obtained, and the optimal slope stability evaluation model after evaluation is used to process it to obtain the evaluation result of slope stability.

2. The method for rapid assessment of slope stability as described in claim 1, characterized in that, The characteristic information includes the unit weight, cohesion, internal friction angle, slope height, slope angle, and pore pressure ratio of the slope material; wherein, the pore pressure ratio is defined as the ratio of the pore water pressure at a certain point in the slope to its vertical self-weight stress.

3. The method for rapid assessment of slope stability as described in claim 1, characterized in that, The process of designing the corresponding numerical simulation scheme includes: ensuring that the mean and standard deviation of each feature parameter in the numerical simulation database and the case dataset are consistent, and that the maximum and minimum values ​​are similar, and then using normal distribution sampling to generate the numerical simulation scheme.

4. The method for rapid assessment of slope stability as described in claim 3, characterized in that, Numerical simulation results are obtained through batch calculations.

5. The method for rapid assessment of slope stability as described in claim 1, characterized in that, The process of constructing a slope stability evaluation model includes: using a probabilistic neural network, which consists of an input layer, a pattern layer, a summation layer, and an output layer. The input layer is responsible for reading learning samples from a large database of slope stability data; the pattern layer is trained on the data features of the learning samples to extract the patterns and relationships contained in the samples, providing a basis for the identification and classification of slope stability; the summation layer performs summation operations on the pattern vectors of each category to complete the probability density estimation of each category; and the output layer determines the final classification result based on the category with the highest probability density. A set of intelligent identification models for slope stability with different combinations of input parameters is generated using probabilistic neural networks.

6. The method for rapid assessment of slope stability as described in claim 5, characterized in that, Each neural network unit in the summation layer is connected to only the corresponding type of model neuron and is estimated according to the Parzen window function method and various types of conditional probability densities.

7. The method for rapid assessment of slope stability as described in claim 1, characterized in that, The process of training a slope stability evaluation model based on an enhanced training set includes training each model generated using a probabilistic neural network using an enhanced training set.

8. The method for rapid assessment of slope stability as described in claim 1, characterized in that, The process of evaluating the performance of the slope stability assessment model using a test set includes selecting the model with the best prediction accuracy. The model with the highest prediction accuracy is selected when considering the characteristics of unit weight, cohesion, internal friction angle, slope angle, slope height and pore pressure ratio. The performance of this model is then evaluated.

9. The method for rapid assessment of slope stability as described in claim 1, characterized in that, When performing performance evaluation, the parameters evaluated include accuracy, precision, recall, and F1-Score.

10. A rapid slope stability assessment system, characterized in that, include: The feature information extraction module is configured to collect slope engineering cases through field exploration or literature review, extract feature information, and establish a case dataset. The data augmentation module is configured to perform statistical analysis on the parameters in the case dataset, design corresponding numerical simulation schemes, obtain numerical simulation results, and build a numerical simulation database to achieve data augmentation. The model building and training module is configured to build a slope stability evaluation model based on a classification neural network algorithm, divide the case dataset into a training set and a test set, integrate a numerical simulation database into the training set, and build an enhanced training set. The slope stability evaluation model is trained based on the enhanced training set, and the performance evaluation of the slope stability evaluation model is completed using the test set. The stability assessment module is configured to acquire the characteristic information of the target slope, process it using the optimal slope stability evaluation model after assessment, and obtain the slope stability assessment result.

Images