Method for observing and warning hydraulically driven landsliding

Abstract

To provide a landsliding observing and warning method, an electronic device, and a storage medium, which are suited for a technical field of landsliding geological disaster prevention control.SOLUTION: This landsliding observing and warning method comprises: acquiring displacement observation data relating to a plurality of observation points of a landsliding body to decompose the displacement observation data into an inclination item and a period item; acquiring a best dynamic parameter at a different time point through reversal in accordance with the inclination items of the observation points; acquiring a safe coefficient time sequence of the sliding body in accordance with the best dynamic parameter at the different time point; combining the respective inclination and period items to acquire a combined inclination item sequence and a combined period item sequence; superimposing the two sequences to acquire a sum deformation time sequence; establishing a sum deformation speed sum deformation speed time sequence in accordance with the sum deformation time sequence; generating the correlation between the safe coefficient time sequence and the sum deformation speed time sequence; establishing at least one sum deformation speed critical value in accordance with the correlation; and executing observation and warning for the landsliding body on the basis of at least one sum deformation speed critical value.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention belongs to the technical field of landslide geological disaster prevention and control, and particularly relates to a landslide observation and warning method, an electronic device, and a storage medium.

Background Art

[0002] Currently, landslide observation has gradually developed from the conventional point-type manual observation to the multi-dimensional collaborative observation of "sky (optical remote sensing and radar) - air (unmanned aerial vehicle photogrammetry) - ground (special observations such as the global positioning satellite system, total station, etc.)". In addition, the observation content has also developed from single deformation observation to the observation of various heterogeneous data such as stress, inclination, groundwater, rainfall, micro-vibration, sound wave, and video. However, the current landslide observation methods and warning models are essentially a simple data processing method, and none of them consider the geological conditions of the landslide body itself and the slope body structure form. Therefore, the geological and mechanical basis of the observation and warning model is insufficient, and the accuracy is not high.

Summary of the Invention

Problems to be Solved by the Invention

[0003] In view of this, in the embodiments of the present invention, a landslide observation and warning method, an electronic device, and a storage medium for improving the accuracy of landslide observation and warning are provided.

Means for Solving the Problems

[0004] In the first aspect of the embodiments of the present invention, obtain displacement observation data of a plurality of observation points of a landslide body, and decompose the displacement observation data into a trend term and a periodic term; according to the trend terms of the plurality of observation points, obtain the best mechanical parameters at different time points by inversion, and evaluate the stability of the landslide body at different time points according to the best mechanical parameters at different time points to obtain the safety factor time series of the landslide body; Fuse the trend terms and periodic terms of multiple observation points respectively to obtain a fused trend term series and a fused periodic term series, superimpose the fused trend term series and the fused periodic term series to obtain a comprehensive deformation time series, and determine a comprehensive deformation rate time series according to the comprehensive deformation time series, and Create a correspondence relationship between the safety factor time series and the comprehensive deformation rate time series, determine at least one comprehensive deformation rate critical value according to the correspondence relationship, and perform observation and warning on the landslide body based on at least one comprehensive deformation rate critical value. A landslide observation and warning method is provided.

[0005] Furthermore, obtaining the best mechanical parameters at different time points by inversion according to the trend terms of multiple observation points is According to the trend terms of multiple observation points, a weighted objective function of the trend terms and mechanical parameters of multiple observation points JPEG2025077974000002.jpg1342 Create it, among which, T(x) Is the objective function value, TIFF2025077974000003.tif611 is the mathematical relationship between the trend term and the mechanical parameter of the i-th observation point, T Di Is the trend term of the i-th observation point, x is the combined vector of mechanical parameters, w i Is the trend term weight value of the i-th observation point, and Calculate the mechanical parameters that minimize the objective function value at different time points, and obtain the best mechanical parameters at different time points.

[0006] Furthermore, the mathematical relationship between the trend term and the mechanical parameter is JPEG2025077974000004.jpg1337, and Among them, JPEG2025077974000005.jpg85 is the fitting coefficient corresponding to the mechanical parameter waiting for inversion of the i-th, M is the number of mechanical parameters waiting for inversion, and JPEG2025077974000006.jpg85 is a preset value, JPEG2025077974000007.jpg85 is the i-th dynamic parameter waiting for inversion.

[0007] Furthermore, according to the best dynamic parameters at different time points, evaluating the stability of the landslide body at different time points and obtaining the time series of the safety factor of the landslide body, includes inputting the best dynamic parameters at different time points into a preset landslide body numerical analysis model, evaluating the stability of the landslide body at different time points by the numerical simulation forward calculation method, and obtaining the time series of the safety factor of the landslide body.

[0008] Furthermore, fusing the trend terms and periodic terms of multiple observation points respectively to obtain a fused trend term series and a fused periodic term series, performing weighted connection on the trend terms of multiple observation points according to the pre-calculated trend term weight values of individual observation points to obtain a fused trend term series, and performing weighted connection on the periodic terms of multiple observation points according to the pre-calculated periodic term weight values of individual observation points to obtain a fused periodic term series.

[0009] Furthermore, the trend term weight value of an individual observation point is calculated according to JPEG2025077974000008.jpg1726, where w i is the trend term weight value of the i-th observation point, n is the number of observation points, H i is the entropy of the i-th observation point, is JPEG2025077974000009.jpg1338, l is the number of trend terms of an individual observation point, is JPEG2025077974000010.jpg1829, and T Di is the trend term of the i-th observation point.

[0010] Furthermore, the periodic term weight value of an individual observation point Obtain the reservoir water level data and rainfall data of each observation point, According to the reservoir water level data and rainfall data, calculate the maximum reservoir water level difference, maximum reservoir water level change rate, moving average rainfall, and effective rainfall within the preset time period of each observation point, After performing the averaging process on the periodic terms and each index of each observation point, calculate the correlation degree between the periodic terms and each index respectively. Each index includes the maximum reservoir water level difference, maximum reservoir water level change rate, moving average rainfall, and effective rainfall, According to the correlation degree, calculate the modulus length of correlation degree corresponding to each observation point, JPEG2025077974000011.jpg1821 Calculate the periodic term weight value of each observation point accordingly, among which, JPEG2025077974000012.jpg64 is the periodic term weight value of the i-th observation point, n is the number of observation points, and Rl i is the modulus length of correlation degree corresponding to the i-th observation point, and is calculated by the method.

[0011] Furthermore, creating the correspondence relationship between the safety factor time series and the comprehensive deformation speed time series, and determining at least one comprehensive deformation speed critical value according to the correspondence relationship, Calculating the grey correlation coefficient between the safety factor and the comprehensive deformation speed, Using the grey correlation coefficient and the safety factor as the input of the random forest model, and the comprehensive deformation speed as the output of the random forest model, and training the comprehensive deformation speed prediction model, Inputting at least one safety factor critical value and the grey correlation coefficient preset value into the comprehensive deformation speed prediction model to obtain at least one comprehensive deformation speed critical value, Furthermore, based on at least one comprehensive deformation speed critical value, observing and warning the landslide body, Observing the real-time comprehensive deformation speed of the landslide body, Determining the landslide risk degree according to the magnitude relationship between the real-time comprehensive deformation speed and at least one comprehensive deformation speed critical value, and performing observation and warning on the landslide body based on the landslide risk degree.

[0012] In a second aspect of an embodiment of the present invention, there is provided an electronic device including a memory, a processor, and a computer program stored in the memory and operable on the processor, wherein when the processor executes the computer program, the steps of the method in the first aspect or any one of the implementation manners of the first aspect are realized.

[0013] In a third aspect of an embodiment of the present invention, there is provided a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the steps of the method in the first aspect or any one of the implementation manners of the first aspect are realized.

Advantages of the Invention

[0014] The embodiments of the present invention have the following beneficial effects compared with the prior art.

[0015] In the embodiments of the present invention, by reversing the trend terms of a plurality of observation points, the best mechanical parameters at different time points are obtained, and further the time series of the safety factor of the landslide body is obtained. Then, by creating the time series of the comprehensive deformation speed, a reasonable comprehensive deformation speed critical value is determined according to the correspondence relationship between the time series of the safety factor and the time series of the comprehensive deformation speed, and observation and warning are performed on the landslide body. That is, in the present solution, the influence of multi-dimensional factors such as the geological conditions, deformation rules, and safety factor of the landslide body on the observation and warning is considered simultaneously, and an alarm model with a geological and mechanical basis is constructed by using the safety factor and the comprehensive deformation speed, avoiding simply analyzing the observation data from a mathematical perspective, and improving the accuracy of the landslide observation and warning.

Brief Description of the Drawings

[0016] The following briefly introduces the drawings that need to be used in the description of the embodiments or the prior art in order to more clearly explain the technical solutions in the embodiments of the present invention.

[0017]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Modes for Carrying Out the Invention

[0018] In response to the deficiencies of the current existing technologies, the present invention comprehensively considers the influence of multi-dimensional factors such as the geological conditions, deformation rules, and safety factors of landslide bodies on landslide body observation and warning, and provides a hydraulic-driven landslide observation and warning method with improved accuracy and timeliness of hydraulic-driven landslide observation and warning. As shown in FIGS. 1 and 2, In step S101, displacement observation data of a plurality of observation points of the landslide body are acquired, and the displacement observation data is decomposed into a trend term and a periodic term.

[0019] At each individual observation point, data can be collected at regular intervals by a displacement sensor. As a result, the obtained displacement observation data is a displacement data series arranged according to time, and the displacement data that has not been decomposed in the displacement data series is called the original total displacement. The empirical mode decomposition method is adopted to decompose the original total displacement into a trend term and a periodic term.

[0020] In step S102, according to the trend terms of a plurality of observation points, inversion is performed to obtain the best mechanical parameters at different time points. According to the best mechanical parameters at different time points, the stability of the landslide body at different time points is evaluated, and a time series of the safety factor of the landslide body is obtained.

[0021] By creating a weighted objective function between the trend terms and mechanical parameters of a plurality of observation points, the best mechanical parameters of the rock and soil body that minimize the objective function value at different time points can be searched by means of a mathematical optimization method.

[0022] Among them, the weighted objective function is as follows: JPEG2025077974000013.jpg1366, JPEG2025077974000014.jpg1337 In the formula, T(x) is the objective function value, JPEG2025077974000015.jpg611 is the mathematical relationship between the trend term and the mechanical parameters of the i-th observation point, T Di is the trend term of the i-th observation point, x is the combined vector of mechanical parameters, w iis the trend term weight value of the i-th observation point, JPEG2025077974000016.jpg85 is the fitting coefficient corresponding to the i-th mechanics parameter waiting for inversion, m is the number of mechanics parameters waiting for inversion, JPEG2025077974000017.jpg85 is the preset value, JPEG2025077974000018.jpg85 is the i-th mechanics parameter waiting for inversion.

[0023] The mathematical relationship between the trend term and the mechanics parameter JPEG2025077974000019.jpg611 takes into account the geological conditions of the landslide body, improving the accuracy of prediction. The above-mentioned mechanics parameters include elastic modulus, Poisson's ratio, adhesion force, friction angle, etc., and the mathematical relationship between the trend term and the mechanics parameter JPEG2025077974000020.jpg611 is obtained by calculation through tests.

[0024] Input the best mechanics parameters at different time points into the preset landslide body numerical analysis model, evaluate the stability of the landslide body at different time points by the numerical simulation forward calculation method, and obtain the time series of the safety factor of the landslide body. The landslide body numerical analysis model is a three-dimensional numerical analysis model. This model individually calls each calculation result file, adopts the strength reduction method to calculate the safety factor at each observation time point, and thereby obtains the time course curve of the safety factor within the entire observation period.

[0025] In step S103, the trend terms and periodic terms of multiple observation points are respectively fused to obtain a fused trend term series and a fused periodic term series, the fused trend term series and the fused periodic term series are superimposed to obtain a comprehensive deformation time series, and a comprehensive deformation velocity time series is determined according to the comprehensive deformation time series. According to the pre-calculated trend term weight values of individual observation points, weighted combination is performed on the trend terms of multiple observation points to obtain a fused trend term series. The expression is as follows: In JPEG2025077974000021.jpg, Equation 980, F TD represents the displacement of the fusion tendency term, and w i is the tendency term weight value of the i-th observation point, and T Di is the tendency term of the measured value of the i-th observation point.

[0026] According to the pre-calculated periodic term weight values of individual observation points, weighted combination is performed on the periodic terms of multiple observation points to obtain a fused periodic term series. The expression is as follows: In JPEG2025077974000022.jpg, Equation 874, F PD represents the displacement of the fused periodic term, and λ i is the periodic term weight value of the i-th observation point, and P Di is the periodic term of the measured value of the i-th observation point.

[0027] F TD and F PD are superimposed to obtain the comprehensive deformation time series Com_D of all observation points. In JPEG2025077974000023.jpg, Equation 635, the gradient of the comprehensive deformation time series, that is, the comprehensive deformation speed, is obtained by performing gradient calculation on the comprehensive deformation time series Com_D by the average gradient method.

[0028] In step S104, a correspondence relationship between the safety factor time series and the comprehensive deformation speed time series is created. According to the correspondence relationship, at least one comprehensive deformation speed critical value is determined, and based on at least one comprehensive deformation speed critical value, an observation warning is given to the landslide body.

[0029] The safety factor can reflect the dangerous situation of the landslide. According to the correspondence relationship between the safety factor time series and the comprehensive deformation speed time series, the comprehensive deformation speed critical values corresponding to several safety factor critical values can be calculated. By observing the comprehensive deformation speed of the landslide body, the risk degree of the landslide body is judged according to the magnitude relationship between the comprehensive deformation speed and each comprehensive deformation speed critical value.

[0030] For example, in this embodiment, the following four-stage progressive warning model is created.

[0031] In Stage I, Com_Vel_Cur < Com_Vel_1.25, and the landslide body is in a stable state.

[0032] In Stage II, Com_Vel_1.25 ≤ Com_Vel_Cur < Com_Vel_1.05, and the landslide body is in a nearly stable state.

[0033] In Stage III, Com_Vel_1.05 ≤ Com_Vel_Cur < Com_Vel_1.0, and the landslide body is in a state of lack of stability.

[0034] In Stage IV, Com_Vel_1.0 ≤ Com_Vel_Cur, and the landslide body is in an unstable state.

[0035] Com_Vel_Cur represents the comprehensive deformation speed of the landslide body at the current time point, and Com_Vel_1.0, Com_Vel_1.05, and Com_Vel_1.25 represent the corresponding comprehensive deformation speeds when the safety factors are equal to 1.0, 1.05, and 1.25 respectively.

[0036] The tendency term weight value w of each observation point i can be calculated by the following method. In Equation 1725 of JPEG2025077974000024.jpg, n is the number of observation points. H i is the entropy of the i-th observation point, and the calculation formula is as follows. In Equation 1338 of JPEG2025077974000025.jpg, l is the number of tendency terms of each observation point. f ji The calculation formula of is as follows. In Equation 1829 of JPEG2025077974000026.jpg, T Di is the tendency term of the i-th observation point. When f ji = 0, f ji ln(f ji ) is treated as 0.

[0037] Periodic term weight value of each observation point JPEG2025077974000027.jpg64 can be calculated by the following method, that is, JPEG2025077974000028.jpg1721, JPEG2025077974000029.jpg64 is the periodic term weight value of the i-th observation point, n is the number of observation points, and Rl i is the correlation modulus length corresponding to the i-th observation point.

[0038] Assuming that the preset time period is one month, the above calculation process of JPEG2025077974000030.jpg64 will be described in detail as follows.

[0039] In step 1, calculate the monthly maximum reservoir water level difference (D_Level), monthly maximum reservoir water level change rate (D_Rate), monthly moving average rainfall (R_Ave), and monthly effective rainfall (R_Val). JPEG2025077974000031.jpg5474 In the formula, W 0 represents the reservoir water level value at the current observation time point, and W i represents the reservoir water level value i days before the current observation time point, and R 0 represents the rainfall value at the current observation time point, and R i represents the rainfall value i days before the current observation time point.

[0040] In step 2, perform an averaging process on the periodic term displacement (P Di ), monthly maximum reservoir water level difference (D_Level), monthly maximum reservoir water level change rate (D_Rate), monthly moving average rainfall (R_Ave), and monthly effective rainfall (R_Val). JPEG2025077974000032.jpg7194 In the formula, JPEG2025077974000033.jpg66, JPEG2025077974000034.jpg617, JPEG2025077974000035.jpg616, JPEG2025077974000036.jpg614, JPEG2025077974000037.jpg613 respectively represent the average values of the P Di , D_Level, D_Rate, R_Ave, and R_Val vectors.

[0041] In step 3, calculate the correlation coefficient between the periodic term of each observation point and the above four indicators, JPEG2025077974000038.jpg69126 where, JPEG2025077974000039.jpg725, JPEG2025077974000040.jpg723, JPEG2025077974000041.jpg721, JPEG2025077974000042.jpg720 respectively represent the correlation coefficient values between the D_Level, D_Rate, R_Ave, and R_Val at time t of the periodic term of the i-th observation point.

[0042] In step 4, calculate the degree of correlation, JPEG2025077974000043.jpg7158 where, JPEG2025077974000044.jpg724, JPEG2025077974000045.jpg723, JPEG2025077974000046.jpg721, JPEG2025077974000047.jpg720 respectively represent the degree of correlation between the periodic term of the i-th observation point and D_Level, D_Rate, R_Ave, and R_Val.

[0043] In step 5, calculate the modulus length of the degree of correlation of each observation point, JPEG2025077974000048.jpg995 where Rl irepresents the correlation degree modulus length of the periodic term at the i-th observation point.

[0044] In step 6, according to Rl i calculate the weight value of the periodic term at the i-th observation point by using a mathematical formula.

[0045] The observation and warning method according to this embodiment comprehensively considers multi-dimensional factors such as the geological conditions, deformation rules, and safety factor of the landslide body, performs observation and warning on the landslide body, and can effectively improve the observation and warning accuracy and timeliness of the water-driven landslide.

[0046] Figure 3 is an engineering geological plan and displacement observation point layout diagram of a certain landslide body. The original observation data of the total displacement, reservoir water level rise and fall, and rainfall of 12 observation points within the landslide perimeter are as shown in Figure 4. The calculation results of the four indicators D_Level, D_Rate, R_Ave, and R_Val according to the original observation data in Figure 3 are as shown in Figure 5.

[0047] According to the topographic map of the landslide body, combine it with geological exploration data such as geological plan, cross-section, and excavation to create its three-dimensional numerical analysis model, as shown in Figure 6.

[0048] The landslide body model is simplified into four parts: (1) crushed stone, silt clay containing boulders, (2) boulders and crushed stone containing silt clay, (3) sliding zone, and (4) sliding bed. The rock and soil bodies of the four parts all adopt the Mohr-Coulomb elastoplastic constitutive model. Since the surface deformation is mainly due to the superposition of the deformation of the landslide body itself and the sliding along the sliding zone, the total of 12 parameters of the elastic modulus E, Poisson's ratio μ, cohesion c, and friction angle φ of (1), (2), and (3) are determined as inversion waiting parameters, and the unit weights of (1), (2), and (3) and all the parameters of (4) adopt the values from the existing geological exploration data. The value ranges and the values of the deterministic parameters are shown in Table 1.

[0049] Table 1 Values of Rock and Soil Body Parameters JPEG2025077974000049.jpg39153

[0050] The individual inversion waiting parameters are regarded as the influencing factors of the orthogonal test. Each individual influencing factor takes three level values (i.e., the maximum value, the intermediate value, and the minimum value). Without considering the interaction between elements, 12 elements and 3 element levels are represented by the L 27 (3 12 ) orthogonal array is selected, and the required number of test runs is 27. The orthogonal test plan is shown in Table 2.

[0051] Table 2 Orthogonal Test Plan Table JPEG2025077974000050.jpg159166

[0052] Substitute the parameters of all the test plans in Table 1 into the three-dimensional numerical analysis model and perform finite difference "forward calculation" according to the natural working conditions. After completing all the "forward calculations", extract the calculated deformation values at the corresponding positions of the 12 observation points in the landslide body of each "forward calculation" plan, and then adopt the multiple regression analysis method to fit the calculated deformation values other than the 9th, 18th, and 27th plans, and create a functional relationship formula between the calculated deformation values of each observation point and the 12 inversion waiting parameters. The fitting results are shown in Table 3.

[0053] Table 3 Fitting Result Table of Calculated Deformation Values of Observation Points and Geotechnical Mechanics Parameters JPEG2025077974000051.jpg190162

[0054] Finally, substitute the values of the 9th, 18th, and 27th plans into JPEG2025077974000052.jpg1230 to obtain the predicted values of the 12 observation points, compare them with the corresponding numerical simulation deformation values for verification. The mean square difference of the obtained 12 observation points is 0.8012. Where x is the geotechnical mechanics parameter combination vector, and x = [E 1 , μ 1 , c 1 , φ 1 , E 2 , μ2 , c 2 , φ 2 , E 3 , μ 3 , c 3 , φ 3 is.

[0055] The empirical mode decomposition method is adopted to decompose the surface deformation observation data of 12 observation points within the landslide area into a trend term and a periodic term as shown in Fig. 7. According to the entropy weight method, the entropy weight values of the trend term displacements of 12 observation points are calculated, that is, w = [0.0570, 0.1328, 0.0842, 0.0707, 0.0562, 0.0696, 0.0709, 0.0721, 0.1112, 0.1326, 0.0737, 0.0688]. According to the functional relationship between the entropy weight value, the calculated deformation value and the mechanical parameters, and the trend term displacement, an optimization objective function is constructed. According to the value range of the mechanical parameters in Table 1, the upper limit value, the lower limit value and the initial value of the planned inversion parameter (Note: the initial value takes the intermediate value of the upper limit value and the lower limit value) are set, and the simulated annealing method is used to optimize the geotechnical mechanical parameters at different time points. As a result, time-course curves of a total of 12 parameters, namely the elastic modulus E, Poisson's ratio μ, cohesion c and friction angle φ in regions (1), (2) and (3) as shown in Fig. 8, are obtained.

[0056] The mechanical parameters obtained by inverse optimization at each observation time are input into a 3D numerical analysis model, and each calculation result file is called individually. The strength reduction method is adopted to calculate the safety factor at each observation time point. As a result, a time-course curve of the safety factor within the entire observation period as shown in Fig. 9 is obtained.

[0057] The trend terms of 12 observation points are weighted to form one time series, that is, the fused trend term F TDTo obtain. Also, the grey correlation degree method is adopted to calculate the periodic term weight values of 12 observation points, that is, λ = [0.083, 0.0823, 0.0818, 0.0823, 0.0832, 0.0837, 0.0839, 0.0840, 0.0840, 0.0839, 0.0839, 0.0838], and the periodic terms of the 12 observation points are weighted to obtain one time series, that is, the fusion periodic term F PD To obtain. The comparison diagrams of the fusion trend term and the fusion periodic term displacement with the trend term and the periodic term of each observation point before fusion are shown in Figure 7.

[0058] F TD and F PD are superimposed to obtain the comprehensive deformation time series Com_D, realizing the fusion of the deformation data of all observation points. At the same time, the comprehensive deformation velocity time series is calculated. The comparison between the comprehensive deformation and the deformation observation data of each observation point before fusion is shown in Figure 10, and the comprehensive deformation velocity time series is shown in Figure 11.

[0059] As shown in Figure 12, the grey correlation degree method is adopted to calculate the correlation coefficient F_Vel_ξ(t) between the safety factor and the comprehensive deformation velocity at each time point. The grey correlation coefficient and the safety factor are used as the decision attribute input items for the random forest model training, and the comprehensive deformation velocity is used as the result output item of the random forest training model. After completing the model training, the safety factor decision attribute vectors {1.0, 1.05, 1.25} and the correlation coefficient decision attribute vectors {1.0, 1.0, 1.0} are substituted into the trained random forest model. The obtained stage-by-stage critical comprehensive deformation velocities of the landslide body are Com_Vel_1.0 = 0.8546, Com_Vel_1.05 = 0.5802, and Com_Vel_1.25 = 0.0809. As can be seen from Figure 11, the comprehensive deformation velocity at the current observation time point (i.e., the last observation time point) is 0.0599. As can be seen from the constructed four-stage warning model, the landslide body is in a stable state at the current time point. Therefore, it is considered that the landslide body is in the stage I warning at the current time. Also, as can be seen from Figure 9, the safety factor obtained by the inversion calculation at the current time of the landslide body is 1.46, which is greater than 1.25. Therefore, the landslide body is in a stable state.

[0060] FIG. 13 is a schematic structural diagram of a landslide observation and warning device according to an embodiment of the present invention. The landslide observation and warning device 130 includes the following components.

[0061] The acquisition module 131 is used to acquire displacement observation data of a plurality of observation points of the landslide body and decompose the displacement observation data into a trend term and a periodic term.

[0062] The inversion evaluation module 132 obtains the best mechanical parameters at different time points by inversion according to the trend terms of a plurality of observation points, evaluates the stability of the landslide body at different time points according to the best mechanical parameters at different time points, and is used to obtain the safety factor time series of the landslide body.

[0063] The fusion module 133 respectively fuses the trend terms and periodic terms of a plurality of observation points to obtain a fused trend term series and a fused periodic term series, superimposes the fused trend term series and the fused periodic term series to obtain a comprehensive deformation time series, and is used to determine a comprehensive deformation speed time series according to the comprehensive deformation time series. The warning module 134 creates a correspondence relationship between the safety factor time series and the comprehensive deformation speed time series, determines at least one comprehensive deformation speed critical value according to the correspondence relationship, and is used to perform observation and warning on the landslide body based on at least one comprehensive deformation speed critical value.

[0064] FIG. 14 is a schematic diagram of an electronic device 140 according to an embodiment of the present invention.

[0065] As shown in FIG. 14, the electronic device 140 of this embodiment includes a processor 141, a memory 142, and a computer program 143 stored in the memory 142 and operable on the processor 141, such as a landslide observation warning program. When the processor 141 executes the computer program 143, the steps in the above-described landslide observation warning method embodiments, such as steps S101 to S104 shown in FIG. 1, are realized. Alternatively, when the processor 141 executes the computer program 143, the functions of the respective modules in the above-described device embodiments, such as the functions of modules 131 to 134 shown in FIG. 13, are realized.

Claims

1. Obtaining displacement observation data of a plurality of observation points of a landslide body, and decomposing the displacement observation data into a trend term and a periodic term; According to the trend terms of the multiple observation points, inversion is performed to obtain best mechanical parameters at different time points, and according to the best mechanical parameters at different time points, stability of the landslide body at different time points is evaluated, and a time series of a safety factor of the landslide body is obtained. respectively fusing the trend terms and periodic terms of the multiple observation points to obtain a fusing trend term series and a fusing periodic term series; superimposing the fusing trend term series and the fusing periodic term series to obtain an integrated deformation time series; and determining an integrated deformation speed time series according to the integrated deformation time series; Creating a correspondence relationship between the safety factor time series and the total deformation speed time series, determining at least one total deformation speed critical value according to the correspondence relationship, and issuing an observation and warning for the landslide body based on the at least one total deformation speed critical value. A landslide observation and warning method comprising the steps of:

2. The above-mentioned method of obtaining the best dynamic parameters at different time points by inversion according to the trend terms of the plurality of observation points includes the steps of: a weighted objective function of the trend terms of the plurality of observation points and dynamic parameters in response to the trend terms of the plurality of observation points; where T(x) is the objective function value, is the mathematical relationship between the trend term and the dynamic parameters of the i-th observation point, and T Di is the trend term for the i-th observation point, x is the combination vector of dynamic parameters, and w i is the trend term weight value of the i-th observation point, Calculating dynamic parameters that minimize the objective function value at different time points, and obtaining the best dynamic parameters at different time points. The landslide observation and warning method according to claim 1 .

3. The mathematical relationship between the trend terms and the kinetic parameters is: and Among them, is the fitting coefficient corresponding to the i-th dynamics parameter waiting to be inverted, m is the number of dynamics parameters waiting to be inverted, is the preset value, is the dynamics parameter of the i-th inversion waiting time, The landslide observation and warning method according to claim 2 .

4. The above-mentioned method includes: evaluating the stability of the landslide body at different time points according to the best mechanical parameters at different time points; and obtaining a time series of the safety factor of the landslide body. inputting the best mechanical parameters at different time points into a preset numerical analysis model of the landslide body, and evaluating the stability of the landslide body at different time points through a numerical simulation forward calculation method to obtain a time series of safety factor of the landslide body; The landslide observation and warning method according to claim 1 .

5. The above-mentioned trend term and period term of the plurality of observation points are respectively merged to obtain a merged trend term series and a merged period term series, performing a weighted combination of the trend terms of the plurality of observation points according to pre-calculated trend term weight values ​​of each observation point to obtain the fused trend term series; performing a weighted combination on the periodic terms of the plurality of observation points according to pre-calculated periodic term weight values ​​of each observation point to obtain the fused periodic term series; The landslide observation and warning method according to any one of claims 1 to 4.

6. The trend term weight value of each observation point is It is calculated according to Among them, w i is the trend term weight value of the i-th observation point, n is the number of observation points, and H i is the entropy of the i-th observation point, where l is the number of trend terms at each observation point. and T Di is the trend term for the i-th observation point, The landslide observation and warning method according to claim 5.

7. The periodic term weight value of each observation point is Obtain reservoir water level data and rainfall data from each observation point, Calculate a maximum reservoir water level difference, a maximum reservoir water level change rate, a moving average rainfall, and an effective rainfall within a preset time period of each observation point according to the reservoir water level data and the rainfall data; After performing an averaging process on the periodic terms and each index of each observation point, a correlation degree between the periodic terms and each index is calculated, and each index includes the maximum reservoir water level difference, the maximum reservoir water level change rate, the moving average rainfall, and the effective rainfall; Calculating a correlation modulus length corresponding to each observation point according to the correlation degree; Calculate the periodic term weights for each observation point according to is the periodic term weight value of the i-th observation point, n is the number of observation points, and Rl i is the correlation modulus length corresponding to the i-th observation point, calculated by the formula: The landslide observation and warning method according to claim 5.

8. The above-mentioned, creating a correspondence relationship between the safety factor time series and the total deformation speed time series, and determining at least one total deformation speed critical value according to the correspondence relationship, Calculating the grey correlation coefficient between the safety factor and the overall deformation speed; The grey correlation coefficient and the safety factor are input to a random forest model, and the overall deformation speed is output from the random forest model, and a total deformation speed prediction model is trained to obtain it; Inputting at least one safety factor critical value and a grey correlation coefficient preset value into the integrated deformation speed prediction model to obtain the at least one integrated deformation speed critical value; The above-mentioned, performing an observation and warning for the landslide body based on the at least one overall deformation speed critical value, Observing the real-time overall deformation speed of the landslide body; determining a landslide risk level according to a magnitude relationship between the real-time total deformation speed and the at least one total deformation speed critical value; and issuing an observation and warning to the landslide body based on the landslide danger level. The landslide observation and warning method according to any one of claims 1 to 4.

9. An electronic device comprising a memory, a processor and a computer program stored in the memory and operable by the processor, the computer program being executed by the processor to implement the steps of the method according to any one of claims 1 to 8.

1. An electronic device comprising:

10. A computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the steps of the method according to any one of claims 1 to 8. A computer-readable storage medium comprising:

Images