Large-scale landslide comprehensive treatment dynamic monitoring system

Abstract

The invention belongs to the technical field of landslide monitoring, and particularly discloses a large-scale landslide comprehensive treatment dynamic monitoring system, which comprises a monitoring landform drawing module, a monitoring network construction module, a data acquisition module, a data analysis module, a data transmission module and a judgment and early warning module, the monitoring network construction module reasonably distributes monitoring points, the data acquisition module obtains landslide parameters, the data analysis module evaluates the landslide state and predicts the displacement value, the data transmission module guarantees data circulation, and the judgment and early warning module gives an alarm in time. All the modules work cooperatively, all-directional monitoring, accurate analysis and timely early warning of the landslide mass are achieved, compared with a traditional scattered monitoring mode, the landslide dynamic state can be mastered more comprehensively and accurately, and a reliable basis is provided for prevention and treatment.

Description

Technical Field

[0001] The invention belongs to the technical field of landslide monitoring, and in particular relates to a dynamic monitoring system for comprehensive management of large-scale landslides. Background Art

[0002] Landslides are a common and extremely harmful geological disaster that poses a serious threat to human life, property and infrastructure construction. Especially in areas with complex terrain and concentrated precipitation, the frequency and degree of damage of landslides are more significant. Ankang is located between the Qinling Mountains and the Daba Mountains. It has a subtropical continental monsoon climate, with hot and humid summers and autumns. The special geographical and climatic conditions make it a high-incidence area for landslide disasters.

[0003] For the treatment of existing landslides, although engineering treatment measures such as crack backfilling, anchor cable (rod) frame beams, anchor cable anti-sliding piles, intercepting drainage ditches and monitoring have been implemented in the early stage, the effect is not ideal. With the heavy rainfall in summer, the sliding surface of the landslide at the anti-sliding pile position moved downward, and new signs of sliding damage appeared in the landslide. The anti-sliding piles, frame beams, and retaining walls were seriously damaged, and the damage to community buildings and roads was aggravated. In terms of monitoring, previous monitoring methods are difficult to meet the needs of dynamic monitoring of landslides, and it is impossible to capture the slight changes of the landslide body in time, which brings difficulties to disaster warning. Moreover, according to the existing landslide monitoring system, it is impossible to carry out multi-faceted warnings for different monitoring parameters. For example, the Chinese patent with publication number CN114894082A discloses a method for early warning of slope deformation landslides based on mines. Through the division of monitoring areas, deformation early warning monitoring can be carried out more accurately in the supervision area, which is convenient for users to carry out key monitoring of key areas according to their own needs. However, its monitoring parameters and early warning types are relatively single and cannot be applied to complex geographical conditions and climatic conditions. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects in the prior art and provide a dynamic monitoring system for comprehensive management of large-scale landslides.

[0005] The present invention provides a large-scale landslide comprehensive management dynamic monitoring system, comprising: The monitoring landform drawing module is used to draw the topographic map of the landslide body and obtain the elevation and longitudinal length information of the landslide body; A monitoring network construction module, which arranges a number of monitoring points based on the topographic map of the landslide body, wherein the monitoring points include uniform monitoring points and special monitoring points, and monitoring devices are arranged at the uniform monitoring points and special monitoring points to form a three-dimensional monitoring network; A data acquisition module, which obtains landslide parameters of uniform monitoring points and special monitoring points based on monitoring devices; A data analysis module is used to analyze and calculate the landslide parameters, obtain multiple evaluation indicators, determine the current state of the landslide body based on each evaluation indicator, and output the displacement prediction value of the next time step based on the landslide parameters of different monitoring points of the landslide body at multiple time steps; Data transmission module: connected to the data acquisition module and the data analysis module respectively, and sending the landslide parameters to the data analysis module based on wireless transmission technology; Judgment and warning module: connected to the data analysis module, wherein a plurality of groups of risk thresholds are preset in the judgment and warning module, and a first warning strategy is set corresponding to each group of risk thresholds; and a plurality of risk levels are preset in the judgment and warning module, and a second warning strategy is set corresponding to each risk level; the judgment and warning module compares the analysis result output by the data analysis module with the preset risk threshold, and if the risk threshold is exceeded, the first warning strategy corresponding to the risk threshold is executed; and the judgment and warning module determines the risk level based on the analysis result output by the data analysis module, and executes the second warning strategy corresponding to the risk level.

[0006] A further solution is to evenly draw elevation lines on the topographic map based on the elevation information of the landslide body, evenly draw longitudinal length lines based on the longitudinal length of the landslide body, use each intersection of the elevation line and the longitudinal length line as a uniform monitoring point, and use the existing anti-slide piles of the landslide body as well as the rear edge and front edge of the landslide body as special monitoring points.

[0007] A further solution is that the monitoring device includes RTK-GPS monitoring equipment, an inclinometer, a pressure sensor, a strain sensor, and a soil moisture sensor; The RTK-GPS monitoring equipment is used to collect the absolute displacement of the surface of each monitoring point; The inclinometer is used to collect the deep displacement of each monitoring point; The pressure sensor and strain sensor are used to collect force and strain data of the existing anti-slide piles; The soil moisture sensor is used to collect the seepage moisture inside the landslide body.

[0008] A further solution is that the data analysis module includes a displacement gradient analysis unit, a displacement difference analysis unit, a change rate ratio analysis unit, and a seepage humidity analysis unit; The displacement gradient analysis unit is used to obtain displacement gradients of multiple monitoring points at the same time step, and the displacement gradients are used to judge the unevenness of deformation of different monitoring points of the landslide body; The displacement difference analysis unit is used to obtain the displacement difference of multiple monitoring points based on the same time step, and the displacement difference is used to judge the relative motion state between the parts of the landslide body; The change rate ratio analysis unit is used to obtain the change rate ratios of multiple monitoring points at the same time step, and the change rate ratios are used to judge the development trend of landslide deformation; The seepage moisture analysis unit is used to determine whether tensile cracks are formed at the rear edge and the front edge of the landslide body.

[0009] A further solution is that the displacement gradient is the change in displacement per unit depth, and the calculation formula is: Displacement gradient = (absolute surface displacement - deep layer displacement) / depth difference, where the depth difference is the vertical distance from the surface of the monitoring point to the deepest monitoring layer; The displacement difference is the difference between the absolute surface displacement and the deep displacement at the same monitoring point at the same time; The change rate ratio is the ratio between the deep displacement change rate and the surface absolute displacement change rate.

[0010] A further solution is that the data analysis module also includes a landslide displacement prediction model for predicting the displacement prediction value of the landslide body in the next time step. The construction process of the landslide displacement prediction model is: Collect the absolute surface displacement and deep displacement of different monitoring points at multiple time steps; Standardize the absolute surface displacement and deep displacement to make the data at different monitoring points and at different times comparable; The standardized surface absolute displacement and deep displacement data are organized into a three-dimensional tensor format (N, T, M), where N is the number of monitoring points, T is the time step, and M is the displacement characteristics of each monitoring point; The data in three-dimensional tensor format is input into the spatiotemporal convolutional neural network for iterative training, and the displacement prediction value of the next time step is output.

[0011] A further solution is that the data analysis module also includes a stress-strain analysis model for determining the stress and strain of the anti-slide pile and judging whether the anti-slide pile meets the strength requirement. The construction process of the stress-strain analysis model is: The pressure data and strain data collected by the pressure sensor and strain sensor at multiple time steps are collected, and technical experts are organized to label the pressure data and strain data to establish the corresponding relationship between the pressure data and strain data and the anti-slide pile strength assessment; after labeling, the pressure data and strain data are sequentially input into the neural network unit for iterative training to obtain the stress-strain analysis model.

[0012] A further solution is that the risk thresholds include: displacement gradient threshold, displacement difference threshold, change rate ratio threshold, displacement prediction threshold, stress strain threshold, seepage humidity threshold; The first warning strategy includes at least two levels of warning.

[0013] A further solution is that the judgment and warning module also includes a risk level determination unit, which determines the landslide risk level based on the displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity at the same time step, and executes a second warning strategy matching the risk level.

[0014] A further solution is that the risk level determination process is: Technical experts use 1-10 points to score the importance of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity based on their professional knowledge and experience, with 10 points being the most important and 1 point being the least important. The scores of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity are recorded as ; Collect expert scoring results and calculate the weights of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity ; ; In the formula, The value range is 1-6, which refers to one of the six parameters: displacement gradient, displacement difference, change rate ratio, predicted displacement value, stress strain, and seepage humidity. Representation parameters The weight of is a parameter The final score, It is the final score of traversing displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity; Will Normalization is performed and converted into dimensionless data between 0 and 1. The normalization formula is: ; In the formula, It represents the original monitoring values ​​of the six parameters: displacement gradient, displacement difference, change rate ratio, predicted displacement value, stress strain, and seepage humidity. is a parameter The value after normalization is between 0 and 1; This parameter The minimum value in the entire data set, used to determine the lower limit benchmark of the data. This parameter The maximum value in the entire data set, used to determine the upper limit benchmark of the data; According to the determined weight , perform weighted summation on the standardized parameters to obtain the comprehensive risk index : ; Based on comprehensive risk indicators The value of determines multiple risk levels: high risk, medium risk, and low risk; The high risk, medium risk and low risk are respectively provided with a second early warning strategy.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The large-scale landslide comprehensive management dynamic monitoring system of the present invention obtains detailed terrain information of the landslide body through the monitoring landform drawing module, providing a basis for the subsequent monitoring point layout. The monitoring network construction module deploys a variety of monitoring devices at uniform monitoring points and special monitoring points, which can comprehensively collect multiple parameters of the landslide body and realize all-round monitoring of the landslide body state. By sending the output results of the data analysis module to the judgment and early warning module, different early warning strategies for different risk thresholds or risk levels are realized, making landslide monitoring more comprehensive and accurate, and further improving the accuracy of early warning, which can minimize casualties and property losses.

[0016] The data analysis module of the present invention performs calculation and analysis through multiple analysis units and models, and performs accurate warning in conjunction with multiple risk thresholds set by the judgment and early warning module. By converting the absolute surface displacement, deep displacement, force and strain data of anti-slide piles, and the internal seepage humidity of the landslide body into displacement gradient, displacement difference, change rate ratio, seepage humidity, landslide displacement prediction value, anti-slide pile force analysis and other parameters, the uneven degree of deformation of the landslide body, the relative motion state of each part and the deformation development trend can be accurately judged. Compared with traditional monitoring methods, the monitoring dimension is wider, and multiple parameters are combined to determine the risk level, thereby improving the credibility of monitoring.

[0017] The present invention has technical experts score the importance of each monitoring parameter, calculate the weight and normalize the data to obtain a comprehensive risk index, thereby determining the risk level of the landslide, providing a scientific basis for decision-making, and enabling relevant departments to take targeted measures according to different risk levels, reasonably allocate resources, and improve the scientificity and effectiveness of landslide control and prevention work. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The following drawings are only used to illustrate and explain the present invention, and are not used to limit the scope of the present invention, wherein: Figure 1 : Schematic diagram of the monitoring network construction of the present invention; Figure 2 : Schematic diagram of the connection of the monitoring system of the present invention; In the figure: 1. Longitudinal length line; 2. Elevation line; 3. Monitoring point; 4. Monitoring landform drawing module; 5. Monitoring network construction module; 6. Monitoring device; 7. Data acquisition module; 8. Data transmission module; 9. Data analysis module; 10. Judgment and early warning module; 11. RTK-GPS monitoring equipment; 12. Inclinometer; 13. Pressure sensor; 14. Strain sensor; 15. Soil moisture sensor; 16. Displacement gradient analysis unit; 17. Displacement difference analysis unit; 18. Change rate ratio analysis unit; 19. Seepage moisture analysis unit; 20. Landslide displacement prediction model; 21. Stress strain analysis model; 22. Risk level determination unit. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution, design method and advantages of the present invention clearer, the present invention is further described in detail by specific embodiments in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0020] like Figure 2 As shown, the present invention provides a large-scale landslide comprehensive management dynamic monitoring system, comprising: The monitoring landform drawing module 4 uses a drone equipped with an optical camera to obtain images of the landslide area and draw a topographic map of the landslide, thereby obtaining information on the elevation and longitudinal length of the landslide; The monitoring network construction module 5 arranges several monitoring points 3 based on the topographic map of the landslide body, such as Figure 1 As shown, on the topographic map, an elevation line 2 is evenly drawn based on the elevation information of the landslide body, and a longitudinal length line 1 is evenly drawn based on the longitudinal length of the landslide body. The intersection of each elevation line 2 and the longitudinal length line 1 is used as a uniform monitoring point, and the existing anti-slide piles of the landslide body and the rear edge and front edge of the landslide body are used as special monitoring points. Monitoring devices 6 are arranged at the uniform monitoring points and special monitoring points to form a three-dimensional monitoring network; wherein, the monitoring device 6 includes an RTK-GPS monitoring device 11, an inclinometer 12, a pressure sensor 13, a strain sensor 14, and a soil moisture sensor 15; The data acquisition module 7 acquires the landslide parameters of the uniform monitoring points and the special monitoring points based on the monitoring device 6. In this embodiment, the RTK-GPS monitoring device 11 is used to collect the absolute surface displacement of each monitoring point 3; the inclinometer 12 is used to collect the deep displacement of each monitoring point 3; the pressure sensor 13 and the strain sensor 14 are used to collect the force and strain data of the existing anti-slide piles; the soil moisture sensor 15 is used to collect the seepage moisture inside the landslide body; A data analysis module 9 is used to analyze and calculate the landslide parameters, obtain multiple evaluation indicators, determine the current state of the landslide body based on each evaluation indicator, and output the displacement prediction value of the next time step based on the landslide parameters of different monitoring points 3 of the landslide body at multiple time steps; The data transmission module 8 is connected to the data acquisition module 7 and the data analysis module 9 respectively, and sends the landslide parameters to the data analysis module 9 based on the wireless transmission technology. In this embodiment, the wireless transmission technology specifically refers to the LoRa technology. Since the landslide monitoring area is large and relatively remote, and the network infrastructure is not perfect, the LoRa wireless module is used to transmit the landslide displacement, seepage humidity and other data collected by the monitoring device 6 to a gateway several kilometers or even tens of kilometers away, and then the gateway is connected to the Internet to transmit to the data analysis module 9; Judgment and warning module 10: connected to the data analysis module 9, wherein a plurality of groups of risk thresholds are preset in the judgment and warning module 10, and a first warning strategy is correspondingly set for each group of risk thresholds; and a plurality of risk levels are preset in the judgment and warning module 10, and a second warning strategy is correspondingly set for each risk level; the judgment and warning module 10 compares the analysis result output by the data analysis module 9 with the preset risk threshold, and executes the first warning strategy corresponding to the risk threshold if the risk threshold is exceeded; and the judgment and warning module 10 determines the risk level based on the analysis result output by the data analysis module 9, and executes the second warning strategy corresponding to the risk level.

[0021] In the above, the data analysis module 9 includes a displacement gradient analysis unit 16, a displacement difference analysis unit 17, a change rate ratio analysis unit 18, and a seepage humidity analysis unit 19; the displacement gradient analysis unit 16 is used to obtain the displacement gradient of multiple identical monitoring points 3 based on the same time step, and the displacement gradient is used to judge the uneven degree of deformation of different monitoring points 3 of the landslide body; the displacement difference analysis unit 17 is used to obtain the displacement difference of multiple identical monitoring points 3 based on the same time step, and the displacement difference is used to judge the relative motion state between the parts of the landslide body; the change rate ratio analysis unit 18 is used to obtain the change rate ratio of multiple identical monitoring points 3 based on the same time step, and the change rate ratio is used to judge the development trend of the deformation of the landslide body; the seepage humidity analysis unit 19 is used to judge whether tensile cracks are formed at the rear edge and the front edge of the landslide body. Among them, the displacement gradient is the change in displacement within a unit depth, and the calculation formula is: Displacement gradient = (absolute surface displacement - deep layer displacement) / depth difference, where the depth difference is the vertical distance from the surface of the monitoring point to the deepest monitoring layer; The displacement difference is the difference between the absolute surface displacement and the deep displacement at the same monitoring point 3 at the same time; the change rate ratio is the ratio between the deep displacement change rate and the surface absolute displacement change rate. When obtaining the deep displacement change rate, an inclinometer tube is pre-buried in the landslide body, and the inclinometer 12 probe moves along the guide groove to measure the inclination angle. The horizontal displacement at different depths is calculated in combination with the inclinometer tube parameters. During continuous monitoring, the displacement difference at adjacent time points is divided by the interval to obtain the deep displacement change amount, and then the ratio of the adjacent change amount difference to the corresponding time interval is calculated to obtain the deep displacement change rate. When obtaining the surface absolute displacement change rate, for multiple surface absolute displacement changes, the ratio of the adjacent change amount difference to the time interval is calculated to obtain the surface absolute displacement change rate.

[0022] In order to reflect the displacement prediction value of the next time step, in this embodiment, the data analysis module 9 also includes a landslide displacement prediction model 20 for predicting the displacement prediction value of the landslide body in the next time step. The construction process of the landslide displacement prediction model 20 is: Collect the absolute surface displacement and deep displacement of different monitoring points 3 at multiple time steps; The absolute surface displacement and deep displacement were standardized to make the data from different monitoring points3 and at different times comparable; The standardized surface absolute displacement and deep displacement data are sorted into a three-dimensional tensor format (N, T, M), where N is the number of monitoring points 3, T is the time step, and M is the displacement feature of each monitoring point 3. In this embodiment, the displacement feature is the surface absolute displacement and the deep displacement; The data in three-dimensional tensor format is input into the spatiotemporal convolutional neural network for iterative training, and the displacement prediction value of the next time step is output; The spatiotemporal convolutional neural network includes: In the input layer, the number of neurons is the same as the dimension of the input data, that is, N×T×M.

[0023] The convolution layer uses a 3×3×3 convolution kernel to simultaneously capture the feature information of adjacent monitoring points 3 in space and adjacent time steps in time. The convolution layer performs a convolution operation on the input data, slides the convolution kernel on the input data, calculates the weighted sum of the local area, and obtains the feature map; Activation function: ReLU function is applied to the output of the convolutional layer as the activation function to introduce nonlinear factors and enhance the expressiveness of the model; The pooling layer, located after the convolutional layer, uses a 2×2 maximum pooling kernel to reduce the size of the feature map by half; The fully connected layer maps the features extracted by the convolutional layer and the pooling layer to the output layer and makes predictions through the fully connected layer.

[0024] Output layer,The output layer outputs the predicted landslide displacement value.

[0025] When monitoring landslides, it is necessary to focus on the stress condition of the existing anti-slide piles and repair them in time. Therefore, in this embodiment, the data analysis module 9 also includes a stress-strain analysis model 21, which is used to determine the stress and strain of the anti-slide piles and determine whether the anti-slide piles meet the strength requirements. The construction process of the stress-strain analysis model 21 is: The pressure data and strain data collected by the pressure sensor 13 and the strain sensor 14 at multiple time steps are collected, and technical experts are organized to annotate the pressure data and strain data, and establish a corresponding relationship between the pressure data and strain data and the anti-sliding pile strength assessment; specifically, the experts analyze the sorted pressure data and strain data based on their professional knowledge and experience to assess whether the anti-sliding pile strength meets the requirements. The judgment result is associated with the corresponding data and annotated to establish a corresponding relationship between the pressure data and strain data and the anti-sliding pile strength assessment. For example, for a set of pressure and strain data, if the expert judges that the anti-sliding pile strength meets the requirements, it is annotated as "meets the strength requirements". After annotating, the pressure data and strain data are sequentially input into the neural network unit for iterative training to obtain the stress-strain analysis model 21. Specifically, in this embodiment, the stress-strain analysis model 21 selects MLP (multi-layer perceptron) for anti-sliding pile stress-strain analysis. In the MLP network, since the input data is pressure data and strain data, the number of neurons in the input layer is 2, and the input layer is responsible for receiving the pre-processed pressure and strain data and passing it to the next layer. Two hidden layers are set. The number of neurons in the first hidden layer is set to 64, and the number of neurons in the second hidden layer is set to 32. This layer further combines and abstracts the features of the output of the previous layer, explores deeper patterns and associations in the data, and provides more effective feature representation for the classification decision of the output layer. The stress and strain strength of the anti-sliding pile is evaluated as meeting the strength requirements and not meeting the strength requirements. Therefore, the number of neurons in the output layer is set to 2. The output layer uses the Softmax activation function to convert the hidden layer output into a probability distribution of each category. Each neuron corresponds to a category, and its output value represents the probability of the anti-sliding pile being in the state of that category.

[0026] In this implementation, the risk thresholds include: displacement gradient threshold, displacement difference threshold, change rate ratio threshold, displacement prediction threshold, stress strain threshold, seepage humidity threshold; the first early warning strategy includes two levels of warning. Among them, the displacement gradient threshold is set to 0.2-0.5mm / m, and the corresponding first early warning strategy is that when the displacement gradient is lower than 0.2mm / m, it prompts to pay close attention and increase the monitoring frequency; when it exceeds 0.5mm / m, it initiates emergency response and organizes personnel evacuation. The displacement difference threshold is set to 3-5mm, and the corresponding first early warning strategy is that when the displacement difference is lower than 3mm, it reminds to strengthen monitoring, and when the displacement difference exceeds 5mm, it implements emergency avoidance measures. The change rate ratio threshold is set to 2-3, and the corresponding first early warning strategy is that when the change rate ratio is less than 3, it arranges professionals to conduct on-site inspections and analyze development trends; when the change rate ratio exceeds 5, it takes engineering measures to control the development of landslides. The displacement prediction threshold is set to 10-20mm; the corresponding first early warning strategy is that when the predicted displacement value is less than 10mm, it prompts to pay close attention and make emergency preparations; when the predicted displacement value is greater than 20mm, it quickly organizes personnel evacuation. The stress strain threshold is set to 0.5-0.8, and the corresponding first warning strategy is that the ratio of the current stress to the design strength of the anti-slide pile is less than 0.5, check the state of the anti-slide pile and evaluate the safety; the ratio of the current stress to the design strength of the anti-slide pile is greater than 0.8, take reinforcement measures to ensure the safety of the project. The seepage humidity threshold is set to 0.,4-0.8, and the corresponding first warning strategy is that the ratio of the humidity in the current landslide area to the humidity in the surrounding stable area is less than 0.4, strengthen the monitoring of the hydrogeological environment around the landslide body, and the ratio of the humidity in the current landslide area to the humidity in the surrounding stable area is greater than 0.8, and take drainage measures immediately to lower the groundwater level.

[0027] In this embodiment, the judgment and early warning module 10 also includes a risk level determination unit 22, which determines the risk level of landslide occurrence based on the displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity at the same time step, and executes a second early warning strategy that matches the risk level. Specifically, the risk level determination process is: technical experts use professional knowledge and experience to score the importance of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity using a score of 1-10, with 10 being the most important and 1 being the least important. The scores of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity are recorded in turn as ; Collect expert scoring results and calculate the weights of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity ; ; In the formula, The value range is 1-6, which refers to one of the six parameters: displacement gradient, displacement difference, change rate ratio, predicted displacement value, stress strain, and seepage humidity. Representation parameters The weight of is a parameter The final score, It is the final score of traversing displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity; Will Normalization is performed and converted into dimensionless data between 0 and 1. The normalization formula is: ; In the formula, It represents the original monitoring values ​​of the six parameters: displacement gradient, displacement difference, change rate ratio, predicted displacement value, stress strain, and seepage humidity. is a parameter The value after normalization is between 0 and 1; This parameter The minimum value in the entire data set, used to determine the lower limit benchmark of the data. This parameter The maximum value in the entire data set, used to determine the upper limit benchmark of the data; According to the determined weight , perform weighted summation on the standardized parameters to obtain the comprehensive risk index : ; Based on comprehensive risk indicators The value of determines multiple risk levels: high risk, medium risk, and low risk; The high risk, medium risk, and low risk are respectively provided with a second early warning strategy. The corresponding second early warning strategy is to issue a warning that the landslide body is in a state of landslide at any time, organize personnel evacuation and take emergency rescue measures. , which is a medium risk. The corresponding second warning strategy is to issue a warning that the landslide body is in an unstable state and needs to be closely monitored and preventive measures taken. The corresponding second early warning strategy is that the internal structure of the landslide body is intact and the possibility of large-scale landslide is low.

[0028] In one embodiment, a panel of 10 experts scored the importance of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity. The results are shown in the following table:

[0029] The weights for calculating displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity are as follows:

[0030] The calculation results of the standardized values ​​of each parameter are shown in the following table, where stress-strain is the ratio of the stress currently borne by the landslide body to the design strength of the anti-slide piles, and seepage humidity is the ratio of the humidity in the current landslide area to the humidity in the surrounding stable area.

[0031]

[0032] The normalized weight of each parameter is multiplied by the corresponding standardized value, and then these products are added together to calculate the composite risk index. The result is 0.489. The calculation process is as follows: Displacement gradient: 0.167×0.6=0.1002 Displacement difference: 0.122×0.412=0.0503 Change rate ratio: 0.144×0.5=0.072 Predicted displacement value: 0.167×0.5=0.0835 Stress strain: 0.183×0.6=0.1098 Seepage humidity: 0.122×0.6=0.0732 Finally add the above results: 0.1002+0.0503+0.072+0.0835+0.1098+0.0732=0.489.

[0033] According to the set risk level classification standards, the landslide is at medium risk.

[0034] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. Dynamic monitoring system for comprehensive management of large landslides, characterized by: include: The monitoring landform drawing module is used to draw the topographic map of the landslide body and obtain the elevation and longitudinal length information of the landslide body; A monitoring network construction module, which arranges a number of monitoring points based on the topographic map of the landslide body, wherein the monitoring points include uniform monitoring points and special monitoring points, and monitoring devices are arranged at the uniform monitoring points and special monitoring points to form a three-dimensional monitoring network; A data acquisition module, which obtains landslide parameters of uniform monitoring points and special monitoring points based on monitoring devices; A data analysis module is used to analyze and calculate the landslide parameters, obtain multiple evaluation indicators, determine the current state of the landslide body based on each evaluation indicator, and output the displacement prediction value of the next time step based on the landslide parameters of different monitoring points of the landslide body at multiple time steps; Data transmission module: connected to the data acquisition module and the data analysis module respectively, and sending the landslide parameters to the data analysis module based on wireless transmission technology; Judgment and warning module: connected to the data analysis module, wherein a plurality of groups of risk thresholds are preset in the judgment and warning module, and a first warning strategy is set corresponding to each group of risk thresholds; and a plurality of risk levels are preset in the judgment and warning module, and a second warning strategy is set corresponding to each risk level; the judgment and warning module compares the analysis result output by the data analysis module with the preset risk threshold, and if the risk threshold is exceeded, the first warning strategy corresponding to the risk threshold is executed; and the judgment and warning module determines the risk level based on the analysis result output by the data analysis module, and executes the second warning strategy corresponding to the risk level.

2. The large-scale landslide comprehensive management dynamic monitoring system according to claim 1 is characterized in that: On the topographic map, elevation lines are evenly drawn based on the elevation information of the landslide body, and longitudinal length lines are evenly drawn based on the longitudinal length of the landslide body. The intersection of each elevation line and the longitudinal length line is used as a uniform monitoring point, and the existing anti-slide piles of the landslide body and the rear edge and front edge of the landslide body are used as special monitoring points.

3. The large-scale landslide comprehensive management dynamic monitoring system according to claim 1 is characterized in that: The monitoring device includes RTK-GPS monitoring equipment, an inclinometer, a pressure sensor, a strain sensor, and a soil moisture sensor; The RTK-GPS monitoring equipment is used to collect the absolute displacement of the surface of each monitoring point; The inclinometer is used to collect the deep displacement of each monitoring point; The pressure sensor and strain sensor are used to collect force and strain data of the existing anti-slide piles; The soil moisture sensor is used to collect the seepage moisture inside the landslide body.

4. The large-scale landslide comprehensive management dynamic monitoring system according to claim 3 is characterized in that: The data analysis module includes a displacement gradient analysis unit, a displacement difference analysis unit, a change rate ratio analysis unit, and a seepage humidity analysis unit; The displacement gradient analysis unit is used to obtain displacement gradients of multiple monitoring points at the same time step, and the displacement gradients are used to judge the unevenness of deformation of different monitoring points of the landslide body; The displacement difference analysis unit is used to obtain the displacement difference of multiple monitoring points based on the same time step, and the displacement difference is used to judge the relative motion state between the parts of the landslide body; The change rate ratio analysis unit is used to obtain the change rate ratios of multiple monitoring points at the same time step, and the change rate ratios are used to judge the development trend of landslide deformation; The seepage moisture analysis unit is used to determine whether tensile cracks are formed at the rear edge and the front edge of the landslide body.

5. The large-scale landslide comprehensive management dynamic monitoring system according to claim 4 is characterized in that: The displacement gradient is the change in displacement per unit depth, and the calculation formula is: Displacement gradient = (absolute surface displacement - deep layer displacement) / depth difference, where the depth difference is the vertical distance from the surface of the monitoring point to the deepest monitoring layer; The displacement difference is the difference between the absolute surface displacement and the deep displacement at the same monitoring point at the same time; The change rate ratio is the ratio between the deep displacement change rate and the surface absolute displacement change rate.

6. The large-scale landslide comprehensive management dynamic monitoring system according to claim 5 is characterized in that: The data analysis module also includes a landslide displacement prediction model, which is used to predict the displacement prediction value of the landslide body in the next time step. The construction process of the landslide displacement prediction model is as follows: Collect the absolute surface displacement and deep displacement of different monitoring points at multiple time steps; Standardize the absolute surface displacement and deep displacement to make the data at different monitoring points and at different times comparable; The standardized surface absolute displacement and deep displacement data are organized into a three-dimensional tensor format (N, T, M), where N is the number of monitoring points, T is the time step, and M is the displacement characteristics of each monitoring point; The data in three-dimensional tensor format is input into the spatiotemporal convolutional neural network for iterative training, and the displacement prediction value of the next time step is output.

7. The large-scale landslide comprehensive management dynamic monitoring system according to claim 6 is characterized in that: The data analysis module also includes a stress-strain analysis model, which is used to determine the stress and strain of the anti-slide pile and judge whether the anti-slide pile meets the strength requirement. The construction process of the stress-strain analysis model is: The pressure data and strain data collected by the pressure sensor and strain sensor at multiple time steps are collected, and technical experts are organized to label the pressure data and strain data to establish the corresponding relationship between the pressure data and strain data and the anti-slide pile strength assessment; after labeling, the pressure data and strain data are sequentially input into the neural network unit for iterative training to obtain the stress-strain analysis model.

8. The large-scale landslide comprehensive management dynamic monitoring system according to claim 7 is characterized in that: The risk thresholds include: displacement gradient threshold, displacement difference threshold, change rate ratio threshold, displacement prediction threshold, stress strain threshold, seepage humidity threshold; The first warning strategy includes at least two levels of warning.

9. The large-scale landslide comprehensive management dynamic monitoring system according to claim 8 is characterized in that: The judgment and warning module also includes a risk level determination unit, which determines the landslide risk level based on the displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity at the same time step, and executes a second warning strategy matching the risk level.

10. The large-scale landslide comprehensive management dynamic monitoring system according to claim 9 is characterized in that: The risk level determination process is as follows: Technical experts use 1 to 10 points to score the importance of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity based on their professional knowledge and experience, with 10 points being the most important and 1 point being the least important. The scores of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity are recorded as ; Collect expert scoring results and calculate the weights of displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity ; ； In the formula, The value range is 1-6, which refers to one of the six parameters: displacement gradient, displacement difference, change rate ratio, predicted displacement value, stress strain, and seepage humidity. Representation parameters The weight of is a parameter The final score, It is the final score of traversing displacement gradient, displacement difference, change rate ratio, displacement prediction value, stress strain, and seepage humidity; Will Normalization is performed and converted into dimensionless data between 0 and 1. The normalization formula is: ； In the formula, It represents the original monitoring values ​​of the six parameters: displacement gradient, displacement difference, change rate ratio, predicted displacement value, stress strain, and seepage humidity. is a parameter The value after normalization is between 0 and 1; This parameter The minimum value in the entire data set, used to determine the lower limit benchmark of the data. This parameter The maximum value in the entire data set, used to determine the upper limit benchmark of the data; According to the determined weight , perform weighted summation on the standardized parameters to obtain the comprehensive risk index : ； Based on comprehensive risk indicators The value of determines multiple risk levels: high risk, medium risk, and low risk; The high risk, medium risk and low risk are respectively provided with a second early warning strategy.

Images