A reservoir bank landslide early warning method based on dynamics and deformation indicators

Abstract

The present invention provides a reservoir bank landslide early warning method based on dynamics and deformation indicators, belonging to the technical field of rock and soil disaster monitoring. The invention improves the calculation of the displacement tangent angle of displacement monitoring data, calculates the inclination angle tangent angle of the inclination angle, uses the improved Welch spectrum conversion to obtain the natural frequency of the on-site vibration data, and constructs a three-element hierarchical early warning method for reservoir bank landslides through the two deformation indicators of displacement and inclination angle and the natural frequency dynamic indicator. This method realizes the pre-slip early warning and risk early warning of reservoir bank landslides, improves the accuracy and timeliness of the early warning, and is more suitable for early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of rock and soil disaster monitoring, and in particular to a reservoir bank landslide early warning method based on dynamics and deformation indicators. Background Art

[0002] While the rapid development of China's hydropower energy base has brought enormous economic benefits to society, the stability of reservoir slopes under the frequent fluctuations of water levels also poses significant safety risks. Reservoir bank slopes are commonly subject to the resurgence of landslides in the accumulation layer under the influence of hydrodynamic forces. The secondary disasters of landslides inducing surge waves after the landslides enter the water can cause casualties and property losses.

[0003] The progression of reservoir landslides from initiation to instability is a dynamic process, primarily influenced by rainfall, cyclical fluctuations in reservoir water levels, and earthquakes. These processes reduce the strength of the landslide soil, alter the gravity field, and change the seepage field. This in turn disrupts the balance between the anti-sliding and downward forces, ultimately leading to landslide slippage and the subsequent development of a landslide disaster. Landslide motion is typically divided into the initial deformation phase, the uniform deformation phase, the accelerated deformation phase, and the accelerated deformation phase, and landslide early warning is based on this process. For reservoir bank landslides, it can be difficult to clearly distinguish between the initial deceleration deformation phase, the uniform deformation phase, the accelerated deformation phase, and the impending deformation phase. Instead, they exhibit a step-like deformation pattern, alternating between accelerated and uniform deformation phases. For reservoir bank traction landslides with step-like deformation, traditional displacement early warning methods based on deformation phases have low accuracy. Decomposing the displacement of step-like landslides into periodic and trend terms often requires methods such as neural networks, which are computationally complex and lack timeliness in real-world landslide monitoring. These methods are therefore unsuitable for early warning. Summary of the Invention

[0004] To address the aforementioned problems in the prior art, the present invention provides a reservoir bank landslide early warning method based on dynamics and deformation indicators. This invention constructs a three-element hierarchical early warning method for reservoir bank landslides using two deformation indicators, displacement and tilt angle, and the natural frequency dynamics indicator. This method provides both impending and risk warnings for reservoir bank landslides, improving the accuracy and timeliness of warnings and making them more suitable for early warning. To achieve the above objectives, the technical solution is as follows:

[0005] The present invention provides a reservoir bank landslide early warning method based on dynamics and deformation indicators, the method comprising the following steps:

[0006] S1. Deploy displacement monitoring equipment, inclinometers and vibration pickups on the landslide;

[0007] S2. Obtain real-time displacement monitoring data according to the displacement monitoring device.

[0008] According to the inclinometer, real-time tilt angle data is obtained.

[0009] According to the vibration pickup, real-time vibration data is obtained;

[0010] S3. Using a data preprocessing model based on the displacement monitoring data, the tilt angle data, and the vibration data, obtain preprocessed displacement monitoring data, preprocessed tilt angle data, and preprocessed vibration data;

[0011] S4. According to the pre-processed displacement monitoring data, an improved displacement tangent angle calculation method is used to obtain the displacement tangent angle.

[0012] According to the pre-processed tilt angle data, the tilt angle tangent angle is obtained by using the tilt angle tangent angle calculation method.

[0013] According to the pre-processed vibration data, an improved Welch algorithm is used to perform spectrum conversion to obtain the natural frequency;

[0014] S5. According to the displacement tangent angle, the tilt angle tangent angle and the natural frequency, a three-element graded warning level result is obtained by using a warning level calculation method.

[0015] Optionally, the displacement monitoring device includes a GNSS sensor and a tensile displacement meter, the inclinometer includes a MEMS acceleration sensor, and the vibration pickup includes a MEMS acceleration sensor and a piezoelectric acceleration sensor.

[0016] Optionally, obtaining real-time displacement monitoring data according to the displacement monitoring device in S2 includes:

[0017] The components of the gravity acceleration vector in the XYZ three-axis coordinate system inside the MEMS acceleration sensor are captured in real time by the MEMS acceleration sensor. The method for calculating the tilt angle of the MEMS acceleration sensor includes:

[0018]

[0019]

[0020] Where α is the change in the landslide inclination angle; is the acceleration vector at the initial moment; X0 is the acceleration value in the X0-axis direction collected by the MEMS acceleration sensor at the initial moment; Y0 is the acceleration value in the Y0-axis direction collected by the MEMS acceleration sensor at the initial moment; Z0 is the acceleration value in the Z0-axis direction collected by the MEMS acceleration sensor at the initial moment; is the acceleration vector at the current moment; X1 is the X1-axis acceleration value collected by the MEMS acceleration sensor at the current moment; Y1 is the Y1-axis acceleration value collected by the MEMS acceleration sensor at the current moment; Z1 is the Z1-axis acceleration value collected by the MEMS acceleration sensor at the current moment.

[0021] Optionally, the data preprocessing method includes: smoothing of displacement data, denoising of tilt angle, Fourier transform of vibration pickup, and peak value picking.

[0022] Optionally, the improved calculation method of the displacement tangent angle in S4 includes:

[0023]

[0024] Where, α i To improve the displacement tangent angle; v * is the displacement change rate in the displacement accumulation stage of the step-type deformation landslide; t i is the time at moment i; S i t i The displacement of time.

[0025] Optionally, the calculation method of the tilt angle tangent angle in S4 includes:

[0026]

[0027] Where, β i is the inclination angle tangent angle; A i is the tilt angle at a certain moment; t i is the time at moment i; is the rate of change of the tilt angle during the deformation accumulation stage of a step landslide.

[0028] Optionally, the step S4 uses an improved Welch algorithm to perform spectrum conversion based on the preprocessed vibration data to obtain the natural frequency, including:

[0029] S41. Collect vibration data of appropriate length and divide it into multiple segments to obtain multiple data segments with the same number of data. The vibration data calculation formula of the data segments includes:

[0030] x i (n)=x(n+iM-M),0≤n≤M,1≤i≤L,n=0,1,…,N-1

[0031] Where x i (n) is the vibration data of length N, L is the number of vibration data segments, M is the number of vibration data segments, i is the number of vibration data segments, and x(n+iM-M) is the value numbered (n+iM-M) in the i-th segment;

[0032] S42. Obtaining the central amplitude of each segment of vibration data and the central amplitude of all the vibration data based on the data of all the data segments, wherein the method includes:

[0033]

[0034] Where, is the average value of all vibration data, A i (n) is the center amplitude of the i-th segment of vibration data, L is the number of vibration data segments, M is the number of vibration data in each segment, and i is the number of vibration data segments;

[0035] S43. Compare the center amplitude of each segment with the center amplitude average value, remove data segments with excessively large center amplitude differences, and obtain valid data segments. The vibration data calculation formula of the valid data segments includes:

[0036]

[0037] Where, The length is L * Effective vibration data, L * is the number of valid vibration data segments, M is the number of vibration data in each segment, i is the number of vibration data segments, and x(n+iM-M) is the value numbered (n+iM-M) in the i-th segment;

[0038] S44, according to the following formula 9 and the vibration data of the valid data segment, obtain the period diagram of the vibration data of the valid data segment,

[0039]

[0040] Where, I i (ω) is the periodogram of the i-th segment vibration data, w(n) is the appropriate window function, U is the normalization factor, The length is L * The effective vibration data, M is the number of vibration data in each segment;

[0041] S45, the periodograms of each segment are approximately considered to be unrelated, and the frequency spectrum of each segment is obtained by calculating the power spectrum estimation, including:

[0042]

[0043] Where, F xx (e jω ) is the frequency spectrum of the vibration data, P xx (e jω ) is the power spectrum of vibration data, I i (ω) is the periodogram of the i-th segment vibration data, L* is the number of valid vibration data segments;

[0044] S46. Obtain a spectrum graph of each segment based on the frequency spectrum of each segment, and determine the peak value of the spectrum graph of each segment to obtain the frequency value with the maximum peak value of each segment, which is the natural frequency value of the landslide.

[0045] Optionally, obtaining a three-element graded warning level result in S5 according to the displacement tangent angle, the tilt angle tangent angle, and the natural frequency by a warning level calculation method includes the following steps:

[0046] S51. The landslide warning level is divided into five levels: Level I, Level II, Level III, Level IV and Level V;

[0047] S52, obtaining a displacement warning level according to the classification of the displacement tangent angle warning levels;

[0048] S53, obtaining a warning level of the tilt angle according to the classification of the tilt angle tangent angle warning level;

[0049] S54. Obtaining a warning level of the natural frequency according to the natural frequency early warning method;

[0050] S55. At any given time point, according to the warning level of the displacement, the warning level of the tilt angle, and the warning level of the natural frequency, a final warning result is obtained through a warning level comparison model.

[0051] Optionally, the calculation method of the warning level comparison model includes:

[0052] D t =max(D(α),D(β),D(F)) (13)

[0053] Where D t is the three-factor graded warning level; D(α) is the displacement tangent angle warning level; D(β) is the tilt angle tangent angle warning level; D(F) is the warning level based on the natural frequency.

[0054] Compared with the prior art, the technical solution of the present invention has at least the following beneficial effects:

[0055] On the one hand, the above scheme uses the step-like failure characteristics of reservoir bank landslides under cyclic water level fluctuations to propose an early warning method based on the displacement tangent angle and the inclination tangent angle during the displacement accumulation stage of step-like deformation landslides, which facilitates the pre-slip warning and risk warning of reservoir bank landslides. On the other hand, the natural frequency of the reservoir bank landslide, which reflects the structural properties of the landslide itself and is obtained based on the improved Welch method, is more suitable for early warning. Finally, a three-element graded early warning method based on displacement, inclination tangent angle, and natural frequency is proposed for reservoir bank traction landslides. Its warning level distribution is more reasonable, the warning accuracy is higher, and the timeliness is faster. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0057] Figure 1 It is a schematic diagram of the process of the present invention

[0058] Figure 2 This is a characteristic diagram of displacement changes of reservoir bank landslides in step-wise failure under water level fluctuations according to an embodiment of the present invention.

[0059] Figure 3 This is a characteristic diagram of the change in the tilt angle of the reservoir bank landslide under the water level cycle fluctuation of the embodiment of the present invention

[0060] Figure 4 This is a diagram showing the early warning levels of displacement tangent angle and tilt angle for step-type landslides according to an embodiment of the present invention.

[0061] Figure 5 This is a flow chart of an early warning method for reservoir bank landslide based on natural frequency according to an embodiment of the present invention.

[0062] Figure 6 This is a diagram of a three-element graded early warning method for reservoir bank landslide according to an embodiment of the present invention.

[0063] Figure 7 This is a schematic diagram of the actual effect of the three-element graded early warning method for reservoir bank landslides according to an embodiment of the present invention.

[0064] Figure 8 Schematic diagram of the process of spectrum conversion using the improved Welch algorithm of the embodiment of the present invention

[0065] Figure 9 4 is a flowchart of a method for calculating a warning level according to an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0067] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0068] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0069] like Figures 1 to 9 As shown, the present invention provides a reservoir bank landslide early warning method based on dynamics and deformation indicators, the early warning method comprising the following steps:

[0070] S1. Deploy displacement monitoring equipment, inclinometers and vibration pickups on the landslide;

[0071] Specifically, the displacement monitoring device includes a GNSS sensor and a tensile displacement meter, the inclinometer includes a MEMS acceleration sensor, and the vibration pickup includes a MEMS acceleration sensor and a piezoelectric acceleration sensor.

[0072] S2. Obtain real-time displacement monitoring data according to the displacement monitoring device.

[0073] According to the inclinometer, real-time tilt angle data is obtained.

[0074] According to the vibration pickup, real-time vibration data is obtained;

[0075] Specifically, the displacement monitoring device obtains real-time displacement monitoring data, including:

[0076] The components of the gravity acceleration vector in the XYZ three-axis coordinate system inside the MEMS acceleration sensor are captured in real time by the MEMS acceleration sensor. The method for calculating the tilt angle of the MEMS acceleration sensor includes:

[0077]

[0078] Where α is the change in the landslide inclination angle; is the acceleration vector at the initial moment; X0 is the acceleration value in the X0-axis direction collected by the MEMS acceleration sensor at the initial moment; Y0 is the acceleration value in the Y0-axis direction collected by the MEMS acceleration sensor at the initial moment; Z0 is the acceleration value in the Z0-axis direction collected by the MEMS acceleration sensor at the initial moment; is the acceleration vector at the current moment; X1 is the X1-axis acceleration value collected by the MEMS acceleration sensor at the current moment; Y1 is the Y1-axis acceleration value collected by the MEMS acceleration sensor at the current moment; Z1 is the Z1-axis acceleration value collected by the MEMS acceleration sensor at the current moment.

[0079] S3. Using a data preprocessing model based on the displacement monitoring data, the tilt angle data, and the vibration data, obtain preprocessed displacement monitoring data, preprocessed tilt angle data, and preprocessed vibration data;

[0080] Specifically, the data preprocessing method includes: smoothing of displacement data, denoising of tilt angle, Fourier transform of vibration pickup, and peak value picking.

[0081] The displacement data is smoothed using non-parametric regression analysis methods - local weighted regression smoothing method and local weighted regression smoothing method. The non-parametric regression analysis method - local weighted regression smoothing method can eliminate the impact of the main low-frequency noise at the landslide site on the displacement equipment and retain the main trend of the displacement data. The local weighted regression smoothing method adjusts the bandwidth parameters of the time series data to balance the smoothness of the fitting and the sensitivity to data changes when there is no preset data to follow a specific distribution form.

[0082] The tilt angle is denoised using the wavelet decomposition denoising method. The wavelet decomposition denoising method resets the low-energy sub-band coefficients to zero to eliminate the interference of high-frequency noise on the tilt angle and retains the important local features of the tilt angle data. At the same time, the tilt angle signal is decomposed into multiple sub-bands of different frequencies, and the noise distribution is judged according to the energy distribution of each sub-band. Then, the signal within the sub-band range is processed by thresholding to achieve the denoising purpose.

[0083] S4. According to the pre-processed displacement monitoring data, an improved displacement tangent angle calculation method is used to obtain the displacement tangent angle.

[0084] According to the pre-processed tilt angle data, the tilt angle tangent angle is obtained by using the tilt angle tangent angle calculation method.

[0085] According to the pre-processed vibration data, an improved Welch algorithm is used to perform spectrum conversion to obtain the natural frequency;

[0086] Specifically, the improved method for calculating the displacement tangent angle includes:

[0087]

[0088] Where, α i To improve the displacement tangent angle; v * is the displacement change rate in the displacement accumulation stage of the step-type deformation landslide; t i is the time at moment i; S i t i The displacement of time.

[0089] Before the landslide becomes unstable, the displacement and displacement change rate caused by water level fluctuations are highly regular. The amplitude and rate of each step-like displacement change are relatively consistent. Each step-like change stage is called the displacement accumulation stage.

[0090] like Figure 2 As shown in the figure, for a step-type landslide affected by reservoir water level, each water level fluctuation causes cracks to develop and displacement to increase. When the accumulated displacement reaches a critical value, the tensile cracks deepen, and the residual anti-sliding force drops to a critical point. At this point, if the landslide continues to be affected by water level fluctuations or other external factors, the displacement may suddenly accelerate, entering the accelerated deformation stage.

[0091] Specifically, the method for calculating the tilt angle tangent angle includes:

[0092]

[0093] Where, β i is the inclination angle tangent angle; A i is the tilt angle at a certain moment; t i is the time at moment i; is the rate of change of the tilt angle during the deformation accumulation stage of a step landslide.

[0094] like Figure 3 As shown in the figure, for step-type landslides affected by reservoir water level, the tilt angle and displacement are both deformation indicators, both increasing in a step-like manner with reservoir water level fluctuations, and both are suitable for landslide early warning. The displacement tangent angle early warning method is also applicable to the tilt angle indicator.

[0095] Specifically, if Figure 8 As shown, in S4, based on the pre-processed vibration data, an improved Welch algorithm is used to perform spectrum conversion to obtain the natural frequency, including:

[0096] S41. Collect vibration data of appropriate length and divide it into multiple segments to obtain multiple data segments with the same number of data. The vibration data calculation formula of the data segments includes:

[0097] x i(n)=x(n+iM-M),0≤n≤M,1≤i≤L,n=0,1,…,N-1

[0098] Where x i (n) is the vibration data of length N, L is the number of vibration data segments, M is the number of vibration data segments, i is the number of vibration data segments, and x(n+iM-M) is the value numbered (n+iM-M) in the i-th segment;

[0099] S42. Obtaining the central amplitude of each segment of vibration data and the central amplitude of all the vibration data based on the data of all the data segments, wherein the method includes:

[0100]

[0101] Where, is the average value of all vibration data, A i (n) is the center amplitude of the i-th segment of vibration data, L is the number of vibration data segments, M is the number of vibration data in each segment, and i is the number of vibration data segments;

[0102] S43. Compare the center amplitude of each segment with the center amplitude average value, remove data segments with excessively large center amplitude differences, and obtain valid data segments. The vibration data calculation formula of the valid data segments includes:

[0103]

[0104] Where, is the valid vibration data with a length of L*, L* is the number of valid vibration data segments, M is the number of vibration data segments, i is the number of vibration data segments, and x(n+iM-M) is the value numbered (n+iM-M) in the i-th segment;

[0105] S44, according to the following formula 9 and the vibration data of the valid data segment, obtain the period diagram of the vibration data of the valid data segment,

[0106]

[0107] Where, I i (ω) is the periodogram of the i-th segment vibration data, w(n) is the appropriate window function, U is the normalization factor, The length is L * The effective vibration data, M is the number of vibration data in each segment;

[0108] S45, the periodograms of each segment are approximately considered to be unrelated, and the frequency spectrum of each segment is obtained by calculating the power spectrum estimation, including:

[0109]

[0110] Where, F xx (e jω ) is the frequency spectrum of the vibration data, P xx (e jω ) is the power spectrum of vibration data, I i (ω) is the periodogram of the i-th segment vibration data, L * is the number of valid vibration data segments;

[0111] S46. Obtain a spectrum graph of each segment based on the frequency spectrum of each segment, and determine the peak value of the spectrum graph of each segment to obtain the frequency value with the maximum peak value of each segment, which is the natural frequency value of the landslide.

[0112] When vibration data acquired at a 500Hz sampling frequency is divided into time periods of 2, 5, 10, and 20 minutes, the vibration spectra generated using the improved Welch algorithm all exhibit a distinct peak. When vibration data from the reservoir bank landslide site is sampled at a 500Hz sampling frequency, the duration should not be less than 2 minutes.

[0113] S5. According to the displacement tangent angle, the tilt angle tangent angle and the natural frequency, a three-element graded warning level result is obtained by using a warning level calculation method.

[0114] Specifically, if Figure 9 As shown, in S5, according to the displacement tangent angle, the tilt angle tangent angle and the natural frequency, a three-element graded warning level result is obtained by a warning level calculation method, including the following steps:

[0115] S51. The landslide warning level is divided into five levels: Level I, Level II, Level III, Level IV and Level V;

[0116] like Figures 6 and 7 As shown in the figure, Level I (red), Level II (orange), Level III (yellow), Level IV (blue), and Level V (green) correspond to the alarm level, alert level, warning level, caution level, and safety level, respectively, representing the decreasing sequence of landslide risk. Among them, Level I warning represents the highest risk level.

[0117] Level I alarm level: There is a high possibility of landslide, various short-term precursor characteristics are obvious, and the probability of occurrence within a few hours or days is very high.

[0118] Level II warning level: There is a high possibility of landslide, with certain macro-precursor characteristics, and a high probability of occurring within a few days or weeks.

[0119] Level III warning level: There is a high possibility of landslide, with certain deformation characteristics, and a high probability of occurring within weeks or months.

[0120] Level IV Warning Level: The possibility of landslide is small, there are certain deformation characteristics, and it is unlikely to occur within a year.

[0121] Safety level V: The possibility of landslide is very small and there is no obvious deformation.

[0122] S52, obtaining a displacement warning level according to the classification of the displacement tangent angle warning levels;

[0123] The displacement tangent angle warning level is divided into five levels from small to large according to the unequally spaced angles.

[0124] S53, obtaining a warning level of the tilt angle according to the classification of the tilt angle tangent angle warning level;

[0125] The basis for the warning level of the tilt angle tangent angle is the same as that of the improved displacement tangent angle.

[0126] like Figure 4 As shown in the figure, when the improved displacement tangent angle and tilt angle warning level is less than 20°, the landslide warning is at the safety level, and when it is greater than 85°, it is at the alarm level. Since the warning level angle range is only 5°, it is difficult to reach the Level II warning level in actual landslides. Usually, the level rises sharply from the Level V safety level to the Level I alarm level, which is not suitable for step-by-step warning of reservoir bank landslides.

[0127] S54. Obtaining a warning level of the natural frequency according to the natural frequency early warning method;

[0128] like Figure 5 As shown in the figure, when the change of the natural frequency value reaches 2Hz or the natural order increases, the warning level can be judged to have reached Level II warning level.

[0129] S55. At any given time point, according to the warning level of the displacement, the warning level of the tilt angle, and the warning level of the natural frequency, a final warning result is obtained through a warning level comparison model.

[0130] The calculation method of the warning level comparison model includes:

[0131] D t =max(D(α),D(β),D(F)) (13)

[0132] Where D t is the three-factor graded warning level; D(α) is the displacement tangent angle warning level; D(β) is the tilt angle tangent angle warning level; D(F) is the warning level based on the natural frequency.

[0133] The present invention provides a reservoir bank landslide early warning method based on dynamics and deformation indicators. The invention constructs a three-element graded early warning method for reservoir bank landslides through two deformation indicators, displacement and tilt angle, and the natural frequency dynamic indicator, thereby realizing the impending early warning and risk warning of reservoir bank landslides, improving the accuracy and timeliness of the warning, and being more suitable for early warning.

[0134] It will be appreciated that the present invention is described by way of the above embodiments and should not be construed as limiting the embodiments of the present invention and the scope of the present invention. It will be appreciated by those skilled in the art that various changes or equivalent replacements may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application fall within the scope protected by the present invention.

Claims

1. A reservoir bank landslide early warning method based on dynamics and deformation indicators, characterized by: The method comprises the following steps: S1. Deploy displacement monitoring equipment, inclinometers and vibration pickups on the landslide; S2. Obtain real-time displacement monitoring data according to the displacement monitoring device. According to the inclinometer, real-time tilt angle data is obtained. According to the vibration pickup, real-time vibration data is obtained; S3. Using a data preprocessing model based on the displacement monitoring data, the tilt angle data, and the vibration data, obtain preprocessed displacement monitoring data, preprocessed tilt angle data, and preprocessed vibration data; S4. According to the pre-processed displacement monitoring data, an improved displacement tangent angle calculation method is used to obtain the displacement tangent angle. According to the pre-processed tilt angle data, the tilt angle tangent angle is obtained by using the tilt angle tangent angle calculation method. According to the pre-processed vibration data, an improved Welch algorithm is used to perform spectrum conversion to obtain the natural frequency; S5. Obtain a three-element graded warning level result by using a warning level calculation method based on the displacement tangent angle, the tilt angle tangent angle, and the natural frequency; Wherein, said S4 includes: S41. Collect vibration data of appropriate length and divide it into multiple segments to obtain multiple data segments with the same number of data. The vibration data calculation formula of the data segments includes: x i (n)=x(n+iM-M),0≤n≤M,1≤i≤L,n=0,1,…,N-1 Where x i (n) is the vibration data of length N, L is the number of vibration data segments, M is the number of vibration data segments, i is the number of vibration data segments, and x(n+iM-M) is the value numbered (n+iM-M) in the i-th segment; S42. Obtaining the central amplitude of each segment of vibration data and the central amplitude of all the vibration data based on the data of all the data segments, wherein the method includes: Where, is the average value of all vibration data, A i (n) is the center amplitude of the i-th segment of vibration data, L is the number of vibration data segments, M is the number of vibration data in each segment, and i is the number of vibration data segments; S43. Compare the center amplitude of each segment with the center amplitude average value, remove data segments with excessively large center amplitude differences, and obtain valid data segments. The vibration data calculation formula of the valid data segments includes: Where, The length is L * Effective vibration data, L * is the number of valid vibration data segments, M is the number of vibration data in each segment, i is the number of vibration data segments, and x(n+iM-M) is the value numbered (n+iM-M) in the i-th segment; S44, according to the following formula 9 and the vibration data of the valid data segment, obtain the period diagram of the vibration data of the valid data segment, Where, I i (ω) is the periodogram of the i-th segment vibration data, w(n) is the appropriate window function, U is the normalization factor, The length is L * The effective vibration data, M is the number of vibration data in each segment; S45, the periodograms of each segment are approximately considered to be unrelated, and the frequency spectrum of each segment is obtained by calculating the power spectrum estimation, including: Where, F xx (e jω ) is the frequency spectrum of the vibration data, P xx (e jω ) is the power spectrum of vibration data, I i (ω) is the periodogram of the i-th segment vibration data, L * is the number of valid vibration data segments; S46. Obtain a spectrum graph of each segment based on the frequency spectrum of each segment, and determine the peak value of the spectrum graph of each segment to obtain the frequency value with the maximum peak value of each segment, which is the natural frequency value of the landslide.

2. The reservoir bank landslide early warning method based on dynamics and deformation indicators according to claim 1 is characterized in that: The displacement monitoring device includes a GNSS sensor and a tensile displacement meter, the inclinometer includes a MEMS acceleration sensor, and the vibration pickup includes a MEMS acceleration sensor and a piezoelectric acceleration sensor.

3. The reservoir bank landslide early warning method based on dynamics and deformation indicators according to claim 2 is characterized in that: The step S2 of obtaining real-time displacement monitoring data according to the displacement monitoring device includes: The components of the gravity acceleration vector in the XYZ three-axis coordinate system inside the MEMS acceleration sensor are captured in real time by the MEMS acceleration sensor. The method for calculating the tilt angle of the MEMS acceleration sensor includes: Where α is the change in the landslide inclination angle; is the acceleration vector at the initial moment; X0 is the acceleration value in the X0-axis direction collected by the MEMS acceleration sensor at the initial moment; Y0 is the acceleration value in the Y0-axis direction collected by the MEMS acceleration sensor at the initial moment; Z0 is the acceleration value in the Z0-axis direction collected by the MEMS acceleration sensor at the initial moment; is the acceleration vector at the current moment; X1 is the X1-axis acceleration value collected by the MEMS acceleration sensor at the current moment; Y1 is the Y1-axis acceleration value collected by the MEMS acceleration sensor at the current moment; Z1 is the Z1-axis acceleration value collected by the MEMS acceleration sensor at the current moment.

4. The reservoir bank landslide early warning method based on dynamics and deformation indicators according to claim 1 is characterized in that: The data preprocessing method includes: smoothing of displacement data, denoising of tilt angle, Fourier transformation of vibration pickup, and peak value picking.

5. The reservoir bank landslide early warning method based on dynamics and deformation indicators according to claim 1 is characterized in that: The improved calculation method of the displacement tangent angle in S4 includes: Where, α i To improve the displacement tangent angle; v * is the displacement change rate in the displacement accumulation stage of the step-type deformation landslide; t i is the time at moment i; S i t i The displacement of time.

6. The reservoir bank landslide early warning method based on dynamics and deformation indicators according to claim 1 is characterized in that: The calculation method of the tilt angle tangent angle in S4 includes: Where, β i is the inclination angle tangent angle; A i is the tilt angle at a certain moment; t i is the time at moment i; is the rate of change of the tilt angle during the deformation accumulation stage of a step landslide.

7. The reservoir bank landslide early warning method based on dynamics and deformation indicators according to claim 1 is characterized in that: In S5, a three-element graded warning level result is obtained according to the displacement tangent angle, the tilt angle tangent angle, and the natural frequency by a warning level calculation method, including the following steps: S51. The landslide warning level is divided into five levels: Level I, Level II, Level III, Level IV and Level V; S52, obtaining a displacement warning level according to the classification of the displacement tangent angle warning levels; S53, obtaining a warning level of the tilt angle according to the classification of the tilt angle tangent angle warning level; S54. Obtaining a warning level of the natural frequency according to the natural frequency early warning method; S55. At any given time point, according to the warning level of the displacement, the warning level of the tilt angle, and the warning level of the natural frequency, a final warning result is obtained through a warning level comparison model.

8. The reservoir bank landslide early warning method based on dynamics and deformation indicators according to claim 7 is characterized in that: The calculation method of the warning level comparison model includes: D t =max(D(α),D(β),D(F)) (13) Where D t is the three-factor graded warning level; D(α) is the displacement tangent angle warning level; D(β) is the tilt angle tangent angle warning level; D(F) is the warning level based on the natural frequency.

Images