A statistical machine learning-based method for predicting slope stability
By combining the GM(1,N) grey prediction model and the XGBoost risk identification model, the problem of insufficient accuracy in slope stability prediction under small sample conditions is solved. This enables the effective characterization of the long-term evolution trend of slope deformation and the multi-source environmental influencing factors, thereby improving the reliability of early warning and the practicality of engineering applications.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NORTHWEST NORMAL UNIVERSITY
- Filing Date
- 2026-02-09
- Publication Date
- 2026-06-02
Smart Images
Figure CN122132955A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a slope stability prediction method based on statistical machine learning, belonging to the field of geological disaster monitoring and early warning technology. Background Technology
[0002] Slope stability prediction and landslide risk early warning are important research areas in mining engineering, transportation engineering, and geological disaster prevention. With the development of monitoring technology, slope deformation monitoring methods are gradually evolving from manual inspection to automation and multi-source monitoring, enabling the acquisition of multi-dimensional monitoring data including displacement, rainfall, temperature, and humidity.
[0003] In existing technologies, slope stability analysis methods mainly include mechanical calculation methods based on limit equilibrium theory and prediction methods based on statistical analysis or statistical machine learning. Methods based on mechanical models rely on complete and accurate geological parameters, making them difficult to apply in real-time under complex geological conditions and long-term dynamic monitoring scenarios. While methods based on statistical machine learning have certain advantages in modeling nonlinear relationships, they typically rely on large amounts of historical sample data, and the prediction results lack the ability to physically interpret the development trend of slope deformation.
[0004] Furthermore, most existing methods only focus on the state identification at a single point in time, failing to fully utilize the historical deformation evolution of slopes and the long-term driving effect of environmental factors on the deformation process. Under conditions of small samples or incomplete data, the prediction stability is insufficient, and the reliability of early warning needs to be improved.
[0005] Therefore, there is an urgent need for a slope stability prediction and early warning method that can simultaneously take into account the ability to predict deformation trends and identify risks under small sample conditions. Summary of the Invention
[0006] The purpose of this invention is to provide a slope stability prediction method based on statistical machine learning, which can solve the problems of insufficient prediction accuracy of existing slope stability prediction methods under small sample conditions and difficulty in simultaneously characterizing the long-term evolution trend of slope deformation and the coupling relationship of multi-source environmental influencing factors.
[0007] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides a slope stability prediction method based on statistical machine learning, comprising: Acquire historical displacement monitoring data of slope monitoring points, and simultaneously acquire environmental impact factor data corresponding to the historical displacement monitoring data; Using historical displacement monitoring data as the main sequence and environmental impact factor data as the driving sequence, a GM(1,N) grey prediction model is constructed to predict the trend of slope displacement over time. Based on the prediction results of the GM(1,N) grey prediction model and the actual monitoring data, the displacement prediction residual is calculated, and trend characteristic parameters for characterizing the deformation evolution state of the slope are extracted from the displacement change process and the displacement prediction residual. By performing trend decomposition and noise reduction on the displacement time series, the information characterizing the long-term evolution characteristics of the slope is made explicit. By fusing trend characteristic parameters with environmental impact factor data, a feature vector for slope stability risk identification is constructed. The feature vectors are input into the XGBoost risk identification model to classify and identify the slope stability status; Based on the identification results of the XGBoost risk identification model and combined with the preset risk threshold rules, the slope stability risk level is determined and corresponding risk warning information is generated.
[0008] In addition to the first aspect, environmental impact factors include rainfall, temperature, humidity, air pressure, and wind speed.
[0009] In conjunction with the first aspect, the GM(1,N) grey prediction model further predicts the trend of slope displacement over time by accumulating and generating the main sequence and driving sequence and establishing a whitening differential equation.
[0010] In conjunction with the first aspect, further trend characteristic parameters include displacement change trend values, statistical characteristics of displacement prediction residuals, and displacement change rate.
[0011] In addition to the first aspect, the trend characteristic parameters further include one or more combinations of displacement acceleration, residual rate of change, and trend duration.
[0012] In conjunction with the first aspect, further integrating trend characteristic parameters with environmental impact factor data includes: splicing or weighted fusion of trend characteristic parameters and environmental impact factor data.
[0013] In addition to the first aspect, the XGBoost risk identification model further achieves non-linear mapping of feature vectors through a gradient boosting tree structure.
[0014] In conjunction with the first aspect, the risk threshold rule is further determined based on a combination of trend characteristic parameters and environmental impact factor data.
[0015] In conjunction with the first aspect, furthermore, slope stability risk levels include stable state, metastable state, and unstable state.
[0016] Secondly, the present invention provides a slope stability prediction system based on statistical machine learning, comprising: The data acquisition module is used to acquire historical displacement monitoring data of slope monitoring points and simultaneously acquire environmental impact factor data corresponding to the historical displacement monitoring data. The trend prediction module is used to construct a GM(1,N) grey prediction model with historical displacement monitoring data as the main sequence and environmental impact factor data as the driving sequence to predict the trend of slope displacement over time. The risk warning module is used to calculate the displacement prediction residual based on the prediction results of the GM(1,N) grey prediction model and actual monitoring data. It extracts trend feature parameters to characterize the slope deformation evolution state from the displacement change process and the displacement prediction residual. By performing trend decomposition and noise reduction on the displacement time series, the information characterizing the long-term evolution characteristics of the slope is made explicit. The trend feature parameters are fused with environmental impact factor data to construct a feature vector for slope stability risk identification. The feature vector is input into the XGBoost risk identification model to classify and identify the slope stability state. Based on the identification results of the XGBoost risk identification model and combined with preset risk threshold rules, the slope stability risk level is determined and corresponding risk warning information is generated.
[0017] Thirdly, the present invention provides a computer device, comprising: Storage medium: used to store computer programs; Processor: for executing the computer program to implement the slope stability prediction method based on statistical machine learning as described in the first aspect.
[0018] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the slope stability prediction method based on statistical machine learning as described in the first aspect.
[0019] Fifthly, the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the slope stability prediction method based on statistical machine learning as described in the first aspect.
[0020] Compared with the prior art, the beneficial effects of the present invention are: The slope stability prediction method based on statistical machine learning provided by this invention introduces a GM(1,N) grey prediction model to model the long-term evolution trend of slope displacement, effectively reducing the impact of random noise and short-term disturbances in the original monitoring data on the risk identification results. It explicitly introduces trend information reflecting the long-term evolution characteristics of the slope into the input space of the statistical machine learning model, solving the problem that existing statistical machine learning methods are insufficient in characterizing slope deformation trends under small sample conditions, thus improving the reliability of risk identification results. Combining multi-source environmental impact factor data, it constructs a feature vector containing displacement trend features, residual features, and environmental factor features, and introduces a gradient-boosting-based statistical machine learning risk identification model, realizing continuous prediction and risk warning of slope stability status, and improving the practicality and operability of engineering applications. Attached Figure Description
[0021] Figure 1 This is a flowchart of the slope stability prediction method based on statistical machine learning provided in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the changes in the original data of the combined displacement in three directions of the slope monitoring point provided in this embodiment of the invention; Figure 3 This is a schematic diagram showing the change of the original data of the combined displacement of the slope monitoring points in three directions after noise reduction processing, as provided in the embodiment of the present invention. Figure 4 This is a schematic diagram of the slope displacement variation trend and prediction output based on the GM(1,N) grey prediction model provided in the embodiment of the present invention. Figure 5 This is a schematic diagram of the time series output of slope stability risk early warning based on the GM(1,N) grey prediction model and the XGBoost risk identification model provided in the embodiments of the present invention. Detailed Implementation
[0022] The technical solution of the present invention will be further described in detail below with reference to specific embodiments.
[0023] Embodiments of the present invention are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. Unless otherwise specified, embodiments of the present invention and the technical features thereof can be combined with each other.
[0024] This invention provides a slope stability prediction method based on statistical machine learning, comprising: Acquire historical displacement monitoring data of slope monitoring points, and simultaneously acquire environmental impact factor data corresponding to the historical displacement monitoring data; Using historical displacement monitoring data as the main sequence and environmental impact factor data as the driving sequence, a GM(1,N) grey prediction model is constructed to predict the trend of slope displacement over time. Based on the prediction results of the GM(1,N) grey prediction model and the actual monitoring data, the displacement prediction residual is calculated, and trend characteristic parameters for characterizing the deformation evolution state of the slope are extracted from the displacement change process and the displacement prediction residual. By performing trend decomposition and noise reduction on the displacement time series, the information characterizing the long-term evolution characteristics of the slope is made explicit. By fusing trend characteristic parameters with environmental impact factor data, a feature vector for slope stability risk identification is constructed. The feature vectors are input into the XGBoost risk identification model to classify and identify the slope stability status; Based on the identification results of the XGBoost risk identification model and combined with the preset risk threshold rules, the slope stability risk level is determined and corresponding risk warning information is generated.
[0025] The slope stability prediction method based on statistical machine learning provided in this invention can solve the problems of insufficient prediction accuracy of existing slope stability prediction methods under small sample conditions and difficulty in simultaneously characterizing the long-term evolution trend of slope deformation and the coupling relationship of multi-source environmental influencing factors. It is applicable to stability prediction and risk warning of geological bodies such as landslides and slopes under long-term monitoring conditions.
[0026] Figure 1 This is a flowchart of the slope stability prediction method based on statistical machine learning provided in this embodiment. This flowchart only shows the logical sequence of the method in this embodiment. Provided there are no conflicts, different flowcharts may be used. Figure 1 Complete the steps shown or described in the order indicated.
[0027] The slope stability prediction method based on statistical machine learning provided in this embodiment can be applied to a terminal and can be executed by a slope stability prediction system based on statistical machine learning. This system can be implemented by software and / or hardware and can be integrated into the terminal, such as any tablet computer or computer device with communication capabilities.
[0028] This invention provides a slope stability prediction method based on statistical machine learning, which specifically includes the following steps: Step 1: Obtain historical displacement monitoring data of slope monitoring points, and simultaneously obtain environmental impact factor data corresponding to the historical displacement monitoring data; In this embodiment, the displacement monitoring data includes the displacement data of the slope in the three directions of X-axis, Y-axis and Z-axis; the environmental influencing factors include rainfall, temperature, humidity, air pressure and wind speed.
[0029] Specifically, such as Figure 2 As shown, multiple monitoring points were selected within the slope area, and displacement data of these points were continuously collected. Figure 3 As shown, the displacement monitoring data is preprocessed, including outlier removal, missing value imputation, and noise reduction. Based on the preprocessed displacement monitoring data in three directions, the combined displacement sequence of the slope is calculated, which serves as the master sequence for the subsequent grey prediction model. Simultaneously, environmental impact factor data for the corresponding time period are acquired, and the collected data are constructed into a slope monitoring time series dataset in chronological order.
[0030] Obtain the displacement data of slope monitoring points over a continuous time series, and denote it as the original displacement series. : ; in, Indicates the first The original displacement value at each moment.
[0031] Simultaneously acquire environmental impact factor data corresponding to the displacement data, assuming the first... The original sequences of environmental impact factors for: ; in, express The One value, This refers to the number of environmental impact factors.
[0032] Step 2: Using historical displacement monitoring data as the main sequence and environmental impact factor data as the driving sequence, construct a GM(1,N) grey prediction model to predict the trend of slope displacement over time. In this embodiment, the GM(1,N) grey prediction model generates a whitening differential equation by accumulating the main sequence and the driving sequence, thereby predicting the trend of slope displacement over time.
[0033] Specifically, using the original displacement sequence As the main sequence, perform an accumulation generation operation (1-AGO) on it to obtain the accumulated generation sequence. : ; in, Indicates the first The cumulative displacement value at each moment.
[0034] Similarly, by performing an accumulation operation on the environmental impact factor sequence, the corresponding driving sequence is obtained. : ; in, express The Values.
[0035] Based on this, a GM(1,N) grey prediction model is constructed, whose whitening differential equation is in the form of: ; in, For development coefficient, For the first The gray action coefficient of each environmental impact factor reflects the comprehensive impact of environmental factors such as temperature and rainfall on the displacement evolution of the monitoring point.
[0036] The whitening differential equation is discretized, and the model parameters are estimated using the least squares method to obtain the model parameter vector. : .
[0037] Based on the model parameters, the predicted values of slope displacement are obtained. : .
[0038] To characterize the changing trend of slope displacement, the predicted values of the original displacement sequence are obtained through inverse accumulation generation operation. : .
[0039] The GM(1,N) grey prediction model's prediction output for the slope's combined displacement trend is as follows: Figure 4 As shown.
[0040] Step 3: Based on the prediction results of the GM(1,N) grey prediction model and the actual monitoring data, calculate the displacement prediction residual, and extract the trend characteristic parameters used to characterize the deformation evolution state of the slope from the displacement change process and the displacement prediction residual. By performing trend decomposition and noise reduction on the displacement time series, the information characterizing the long-term evolution characteristics of the slope is made explicit. In this embodiment, the trend characteristic parameters include displacement change trend value, statistical characteristics of displacement prediction residual, and displacement change rate.
[0041] In one possible embodiment, the trend characteristic parameters may also include one or more combinations of displacement acceleration, residual rate of change, and trend duration.
[0042] Specifically, the predicted values of the grey prediction model are compared with the actual monitored values, and the residual sequence is calculated. : .
[0043] In this embodiment, by performing statistical analysis on the residual sequence, characteristic parameters for characterizing the slope deformation trend are extracted, including the residual mean. and residual variance .
[0044] Simultaneously, the displacement change rate is calculated based on the displacement changes at adjacent monitoring times. : ; in, This represents the time interval between adjacent monitoring moments. The trend characteristic parameter is composed of residual statistical characteristics and the rate of displacement change.
[0045] Step 4: Integrate the trend characteristic parameters with the environmental impact factor data to construct a feature vector for slope stability risk identification; In this embodiment, fusing trend feature parameters with environmental impact factor data specifically includes: splicing or weighting the trend feature parameters with environmental impact factor data.
[0046] Eigenvectors are used to characterize the deformation state of slopes at different time scales and their comprehensive characteristics under the influence of environmental factors.
[0047] Step 5: Input the feature vector into the XGBoost risk identification model to classify and identify the slope stability status; In this embodiment, the XGBoost risk identification model uses a gradient boosting tree structure to achieve nonlinear mapping of feature vectors and outputs slope stability risk categories.
[0048] In one possible implementation, the XGBoost risk identification model can be replaced with other supervised learning-based classification models that, after being trained on historical samples, can output the risk level corresponding to slope stability. Based on the risk level, corresponding risk warning information is generated to guide daily slope monitoring, risk management, and engineering decisions.
[0049] Step 6: Based on the identification results of the XGBoost risk identification model and combined with the preset risk threshold rules, determine the slope stability risk level and generate corresponding risk warning information.
[0050] In this embodiment, the risk threshold rule is determined based on a combination of trend characteristic parameters and environmental impact factor data.
[0051] Slope stability risk levels include stable, metastable, and unstable states, which are used to characterize different stages of slope evolution from stability to instability.
[0052] The risk warning output time series obtained by inputting the trend features extracted based on the GM(1,N) grey prediction model and multi-source environmental factors into the XGBoost risk identification model is as follows: Figure 5 As shown. By Figure 5 It can be seen that the risk level of the slope changes dynamically over different time periods. When the risk level output by the model reaches the preset threshold, the corresponding level of early warning information can be triggered.
[0053] This invention provides a slope stability prediction system based on statistical machine learning, comprising: The data acquisition module is used to acquire historical displacement monitoring data of slope monitoring points and simultaneously acquire environmental impact factor data corresponding to the historical displacement monitoring data. The trend prediction module is used to construct a GM(1,N) grey prediction model with historical displacement monitoring data as the main sequence and environmental impact factor data as the driving sequence to predict the trend of slope displacement over time. The risk warning module is used to calculate the displacement prediction residual based on the prediction results of the GM(1,N) grey prediction model and actual monitoring data. It extracts trend feature parameters to characterize the slope deformation evolution state from the displacement change process and the displacement prediction residual. By performing trend decomposition and noise reduction on the displacement time series, the information characterizing the long-term evolution characteristics of the slope is made explicit. The trend feature parameters are fused with environmental impact factor data to construct a feature vector for slope stability risk identification. The feature vector is input into the XGBoost risk identification model to classify and identify the slope stability state. Based on the identification results of the XGBoost risk identification model and combined with preset risk threshold rules, the slope stability risk level is determined and corresponding risk warning information is generated.
[0054] The slope stability prediction system based on statistical machine learning provided in this embodiment of the invention can execute the slope stability prediction method based on statistical machine learning provided in this embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0055] This invention provides a computer device, comprising: Storage medium: used to store computer programs; Processor: Used to execute computer programs to implement the slope stability prediction method based on statistical machine learning provided in the embodiments of the present invention.
[0056] This invention provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the slope stability prediction method based on statistical machine learning provided in this invention.
[0057] This invention provides a computer program product, including a computer program that, when executed by a processor, implements the slope stability prediction method based on statistical machine learning provided in this invention.
[0058] 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 a computer program product embodied 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.
[0059] 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, as well as 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. Figure 1 One or more processes and / or boxes Figure 1 A system that specifies functions in one or more boxes.
[0060] 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 an instruction set implemented in a process. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0061] 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.
[0062] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A slope stability prediction method based on statistical machine learning, characterized in that, include: Acquire historical displacement monitoring data of slope monitoring points, and simultaneously acquire environmental impact factor data corresponding to the historical displacement monitoring data; Using historical displacement monitoring data as the main sequence and environmental impact factor data as the driving sequence, a GM(1,N) grey prediction model is constructed to predict the trend of slope displacement over time. Based on the prediction results of the GM(1,N) grey prediction model and the actual monitoring data, the displacement prediction residual is calculated, and trend characteristic parameters for characterizing the deformation evolution state of the slope are extracted from the displacement change process and the displacement prediction residual. By performing trend decomposition and noise reduction on the displacement time series, the information characterizing the long-term evolution characteristics of the slope is made explicit. By fusing trend characteristic parameters with environmental impact factor data, a feature vector for slope stability risk identification is constructed. The feature vectors are input into the XGBoost risk identification model to classify and identify the slope stability status; Based on the identification results of the XGBoost risk identification model and combined with the preset risk threshold rules, the slope stability risk level is determined and corresponding risk warning information is generated.
2. The slope stability prediction method based on statistical machine learning according to claim 1, characterized in that, Environmental impact factors include rainfall, temperature, humidity, air pressure, and wind speed.
3. The slope stability prediction method based on statistical machine learning according to claim 1, characterized in that, The GM(1,N) grey prediction model predicts the trend of slope displacement over time by accumulating the main sequence and driving sequence and establishing a whitening differential equation.
4. The slope stability prediction method based on statistical machine learning according to claim 1, characterized in that, Trend characteristic parameters include displacement change trend value, statistical characteristics of displacement prediction residual, and displacement change rate.
5. The slope stability prediction method based on statistical machine learning according to claim 4, characterized in that, Trend characteristic parameters also include one or more combinations of displacement acceleration, residual rate of change, and trend duration.
6. The slope stability prediction method based on statistical machine learning according to claim 1, characterized in that, Integrating trend characteristic parameters with environmental impact factor data includes: splicing or weighted fusion of trend characteristic parameters and environmental impact factor data.
7. The slope stability prediction method based on statistical machine learning according to claim 1, characterized in that, The XGBoost risk identification model achieves non-linear mapping of feature vectors through a gradient boosting tree structure.
8. The slope stability prediction method based on statistical machine learning according to claim 1, characterized in that, The risk threshold rule is determined based on a combination of trend characteristic parameters and environmental impact factor data.
9. The slope stability prediction method based on statistical machine learning according to claim 1, characterized in that, Slope stability risk levels include stable state, metastable state, and unstable state.
10. A slope stability prediction system based on statistical machine learning, characterized in that, include: The data acquisition module is used to acquire historical displacement monitoring data of slope monitoring points and simultaneously acquire environmental impact factor data corresponding to the historical displacement monitoring data. The trend prediction module is used to construct a GM(1,N) grey prediction model with historical displacement monitoring data as the main sequence and environmental impact factor data as the driving sequence to predict the trend of slope displacement over time. The risk warning module is used to calculate the displacement prediction residual based on the prediction results of the GM(1,N) grey prediction model and the actual monitoring data. It also extracts trend feature parameters from the displacement change process and the displacement prediction residual to characterize the deformation evolution state of the slope. By performing trend decomposition and noise reduction on the displacement time series, the information characterizing the long-term evolution characteristics of the slope is made explicit. The trend feature parameters are fused with environmental impact factor data to construct a feature vector for slope stability risk identification. The feature vector is input into the XGBoost risk identification model to classify and identify the slope stability status. Based on the identification results of the XGBoost risk identification model and combined with the preset risk threshold rules, the slope stability risk level is determined and corresponding risk warning information is generated.