A probability early warning method based on random fluctuation of landslide deformation monitoring curve

Abstract

The present application relates to the technical field of landslide monitoring, and discloses a probability early warning method based on random fluctuation of a landslide deformation monitoring curve, which comprises the following steps: obtaining landslide deformation monitoring data; establishing a cumulative displacement-time (S-t) curve to calculate the average deformation rate of a uniform deformation stage and the deformation rate v of adjacent two measurements within an equal time interval i ; processing the dimensions of the longitudinal and transverse coordinates of the S-t curve in a consistent manner; if the improved deformation rate conforms to a normal distribution, then calculating the mean value μ v and the standard deviation σ v of the improved rate within the equal time interval; calculating a reliability index β within the equal time interval; calculating a failure probability according to the reliability index, and dividing the landslide early warning level according to the failure probability. In the present application, the randomness of the monitoring data is monitored, and then the probability of landslide instability is reflected, so that the purpose of landslide monitoring and early warning is achieved. The present application avoids the complicated data processing of the traditional method, and has higher reliability and accuracy of early warning and prediction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of landslide monitoring, in particular to a probability early warning method based on random fluctuation of landslide deformation monitoring curve. BACKGROUND

[0002] Landslide monitoring and early warning is an important part of landslide disaster prevention. In the process of landslide monitoring and early warning, efficient monitoring data processing and early warning model are the key to determine whether the early warning is successful. In reality, the landslide monitoring curve has random characteristics. Due to the influence of equipment, environment and many other factors, random fluctuation inevitably occurs in the monitoring curve. In the past, the random fluctuation was regarded as interference noise, and a filtering method was used to remove the noise influence, so as to obtain a relatively smooth monitoring curve. Since the filtering process has a greater dependence on the algorithm, and the human subjective setting of the filtering algorithm parameters is greater, the over-conservative parameter setting may filter out a large amount of effective information, while the over-loose filtering parameter often cannot effectively remove the noise interference, thereby leading to a higher false alarm and missed alarm in the real early warning and prediction. SUMMARY

[0003] The present application aims to provide a probability early warning method based on random fluctuation of landslide deformation monitoring curve, in order to solve the problem that the early warning and prediction after filtering the monitoring data will have a higher false alarm or missed alarm.

[0004] To achieve the above purpose, the present application adopts the following technical scheme: a probability early warning method based on random fluctuation of landslide deformation monitoring curve, comprising the following steps:

[0005] Step 1: Obtain landslide deformation monitoring data, and establish a cumulative displacement-time curve, i.e. S-t curve;

[0006] Step 2: Determine the uniform deformation stage of the S-t curve, and divide the cumulative time displacement curve into several equal time interval sections;

[0007] Step 3: Calculate the average deformation rate of the uniform deformation stage and the deformation rate v i of adjacent two measurements in the equal time interval;

[0008] Step 4: Calculate the improved deformation rate v i / of adjacent two measurements in the equal time interval ;

[0009] Step 5: Process the dimension of the longitudinal and horizontal coordinates of the S-t curve to be consistent, i.e. improve the original S-t curve to T-t curve, where T represents the ratio of the current deformation rate to the uniform deformation stage rate;

[0010] Step 6: If the improved deformation rate conforms to normal distribution, calculate the mean deformation rate μ of the improved deformation rate in the adjacent equal time interval v and the standard deviation σ v ;

[0011] Step 7: Calculate the reliability index β in the equal time interval;

[0012] Step 8: Calculate the failure probability according to the reliability index, and divide the landslide warning level according to the failure probability.

[0013] The applicant finds that the instability change process of the landslide is a development process from disorder to order. In the stable development stage of the landslide, the motion of each monitoring point is in a chaotic state and has strong randomness. With the increase of the deformation and damage degree of the landslide, the control of the overall motion trend of the landslide on each monitoring point on the landslide body is stronger and stronger, and the randomness of each monitoring point is significantly reduced. Therefore, the random fluctuation of the monitoring data can often reflect the stability state of the slope body to a great extent. The scheme considers the randomness of the landslide monitoring curve, reflects the failure probability of the landslide through the random change of the monitoring data, and improves the reliability and accuracy of the early warning and prediction.

[0014] Preferably, step 1 selects a preset equal time interval, and the equal time interval selects the monitoring point data in the time period. In this way, there are enough monitoring points in the time period to have statistical rules.

[0015] Preferably, in step 8, when the failure probability is greater than or equal to 90%, it is a red warning; when the failure probability is greater than or equal to 60% and less than 90%, it is an orange warning; when the failure probability is greater than or equal to 30% and less than 60%, it is a yellow warning; and when the failure probability is less than 30%, it is a blue warning. Among them, the blue warning level is the lowest, and the red warning level is the highest. The landslide failure probability is graded for early warning, different safety measures can be developed for different levels, which facilitates the staff to carry out hierarchical management and protects the life and property safety of the corresponding personnel.

[0016] Preferably, the calculation formula of the failure probability is:

[0017] , wherein: P f represents the failure probability, φ represents the normal distribution cumulative probability distribution function, and β is the reliability index.

[0018] Preferably, the calculation formula of the reliability index β is:

[0019] , wherein, are the mean and standard deviation of the improved deformation rate in a certain equal time interval, respectively, respectively represent the improved mean and standard deviation of deformation rate at the moment of landslide instability according to case statistics, Z is a reliability function function of landslide instability, and is specifically represented as: , v and V respectively represent the improved deformation rate measurement value within a certain equal time interval, V represents the improved rate value at the moment of landslide instability, and v and V are both random variables subject to normal distribution. The applicant finds that the improved rate approximately satisfies the normal distribution characteristics according to a large amount of landslide (slope) instability data, and therefore the mean and standard deviation of the improved rate at the moment of instability are and .

[0020] Preferably, the value and standard deviation of the deformation rate calculated from the measurement data in the selected time period are calculated according to the following formula: , , wherein μ v , σ v respectively represent the improved mean and standard deviation of deformation rate within a certain equal time interval, and v i is the improved deformation rate of adjacent two measurements within a certain equal time interval.

[0021] When the random fluctuations of the monitoring curve are subject to normal distribution, the value and standard deviation of the deformation rate are calculated according to the above formula.

[0022] Preferably, the turning point of the uniform deformation stage and the accelerated deformation stage is determined by the 3σ criterion in determining the uniform deformation stage in step 2.

[0023] Preferably, the n-th monitoring data is considered as the starting point of the accelerated deformation stage when the following formula is met by the continuous three monitoring data in step 2. , wherein: v n+1 , v n+2 , and v n+3 respectively represent the continuous three monitoring data after the n-th data point in the S-t curve, represents the mean value of the rate calculated from the first n data points, represents the standard deviation of the rate calculated from the first n data points.

[0024] Preferably, the tangent slope of the original S-t curve is improved to the tangent slope of the T-t curve in step 5, and the specific steps are as follows:

[0025]

[0026] , wherein: ΔT represents the ratio of the deformation increment of adjacent two measurements to the mean value of the uniform deformation stage, ΔS represents the deformation increment obtained by adjacent two measurements, y i , and y i-1 respectively represent the deformation values of adjacent two measurements, is the deformation rate of the uniform deformation stage;

[0027] The corresponding improved tangent slope can be expressed as:

[0028]

[0029] In the formula, △t i is the time interval between adjacent measurements, v i is the improved deformation rate of adjacent measurements, y i , y i-1 respectively represent the deformation values of adjacent measurements.

[0030] Correspondingly, the improved tangent angle can be expressed as:

[0031] .

[0032] Preferably, the deformation rate v i of adjacent measurements in the equal time interval in step 4 is calculated as: In the formula, △t is the time interval between adjacent monitoring data in the selected equal time interval.

[0033] In the scheme, the deformation rate of each monitoring data in the selected equal time interval is obtained, the random change of the monitoring data is performed, and then the probability of landslide instability is reflected, so that the purpose of landslide monitoring and early warning is achieved; in the traditional monitoring mode, the random fluctuation of the monitoring curve is regarded as noise interference and is removed, and the present application is contrary to the traditional mode, and the random fluctuation is utilized to monitor the stability of the slope body, the complicated data processing of the traditional method is avoided, and the reliability and accuracy of early warning and prediction are higher. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 is the flowchart of the embodiment of the present application.

[0035] Figure 2 is the schematic diagram of the random change of the monitoring data S-t curve in the embodiment of the present application.

[0036] Figure 3 is the improved tangent angle histogram of the example landslide instability in the embodiment of the present application.

[0037] Figure 4 is the improved rate histogram of the example landslide instability in the embodiment of the present application.

[0038] Figure 5 is the example landslide data time displacement result in the embodiment of the present application.

[0039] Figure 6A comparison chart of landslide instability probability and deformation process change for an example embodiment of the present application. DETAILED DESCRIPTION

[0040] Further details are described below through a specific embodiment:

[0041] Embodiment:

[0042] Applicant found through a large number of studies that the instability change process of landslide is a development process from disorder to order. In the stable development stage of landslide, the motion of each monitoring point is in a chaotic state with strong randomness. With the increase of the deformation and damage degree of landslide, the control of the overall motion trend of the landslide on each monitoring point on the landslide body becomes stronger and the randomness of each monitoring point significantly decreases. Therefore, the random fluctuations of the monitoring data can often reflect the stability state of the slope body to a great extent. Applicant overcomes the prejudice that random fluctuations are harmful data in the traditional landslide monitoring method. Based on the above concept, considering the randomness of the landslide monitoring curve, the probability of landslide instability is reflected through the randomness change of the monitoring data. The specific method is as follows:

[0043] A probability early warning method based on random fluctuations of landslide deformation monitoring curve, as shown in Figure 1 , includes the following steps:

[0044] Step 1: Obtain landslide deformation monitoring data and perform preprocessing.

[0045] In this step, a preset equal time interval, such as one day or one week, is selected; and the monitoring point data in the time interval is selected. In this way, there are enough monitoring points in the time interval to have statistical rules. In this step, the actual time interval between adjacent two times of collection of the monitoring point is automatically selected as the preset equal time interval.

[0046] Step 2: Establish a cumulative displacement-time curve, i.e. S-t curve, and divide the cumulative displacement-time curve into several equal time interval segments.

[0047] The cumulative displacement-time curve established in this step is as shown in Figure 2 . The calculation time interval is taken. It is assumed that the random fluctuations of the monitoring data in the fixed time interval conform to the normal distribution. There are N monitoring data in each equal time interval. The monitoring value of a single monitoring data is y0, y1, …, yN-1. Set Δt as the time interval of adjacent two monitoring data in the selected equal time interval. The change rate of adjacent two measurement data is represented as: n

[0048] (1)

[0049] ​The determination of the turning point between the uniform deformation stage and the accelerated deformation stage is the key to determine whether the deformation rate of the uniform deformation stage is accurate. In this step, the turning point between the uniform deformation stage and the accelerated deformation stage is determined by using the 3σ criterion. Specifically, when

[0050] The continuous three monitoring data meet the following requirements, , the nth monitoring data is considered as the starting point of the accelerated deformation stage. In the formula, v n+1 , v n+2 , v n+3 respectively represent the continuous three monitoring data after the nth data point in the S-t curve, , the mean value of the rate calculated from the first n data points, , the standard deviation of the rate calculated from the first n data points.

[0051] Step 3: Calculate the average deformation rate of the uniform deformation stage and the change rate v i。

[0052] In this step, after determining the uniform deformation stage, the average deformation rate v and the change rate v i of the adjacent two measurement data in the same time interval are solved according to the previous rate solving formula (1).

[0053] Step 4: Calculate the improved change rate v i / of the adjacent two measurement data in the same time interval.

[0054] Step 5: Process the dimensions of the longitudinal and horizontal coordinates of the S-t curve, that is, improve the tangent slope of the original S-t curve to the tangent slope of the T-t curve.

[0055] In this step, the tangent slope of the original S-t curve is improved to the tangent slope of the T-t curve. The steps are as follows:

[0056] (2)

[0057] In the formula, △T represents the ratio of the deformation increment of the adjacent two measurements to the average value of the uniform deformation stage, △S represents the deformation increment obtained by the adjacent two measurements, y i , y i-1 respectively represent the deformation values of the adjacent two measurements, is the deformation rate of the uniform deformation stage;

[0058] The corresponding improved tangent slope can be represented as:

[0059] (3)

[0060] wherein, is the improved tangent angle;△t i is the time interval between adjacent two measurements, v i is the improved deformation rate between adjacent two measurements, and the rest of the parameters have the same meaning as before.

[0061] Correspondingly, the improved tangent angle can be expressed as:

[0062] (4).

[0063] Applicants found that if the longitudinal coordinate or the horizontal coordinate is stretched or scaled, i.e., different coordinate scales (unit length representing displacement) are used to make different S-t curves, the S-t curve can still reflect the change characteristics of the initial deformation stage, the constant (uniform) deformation stage and the accelerated deformation stage of the landslide, but the displacement tangent angle of the same monitoring point will change, which leads to uncertainty of the deformation rate between adjacent two measurements, which will affect the result of landslide prediction. The present application overcomes the randomness of the tangent angle caused by the inconsistency of the dimensions of the monitoring data by processing the longitudinal and horizontal coordinates of the S-t curve to be consistent in dimension.

[0064] Step 6: If the improved deformation rate conforms to the normal distribution, calculate the mean μ v and the standard deviation σ v。

[0065] In this step, the tangent slope of the original S-t curve is improved to the tangent slope of the T-t curve, and then the improved deformation rate between adjacent two measurements is calculated. The deformation rate between adjacent two measurements is calculated by formula (3).

[0066] Applicants found that the improved rate approximately meets the normal distribution characteristics according to a large amount of landslide (slope) instability data, so the mean and the standard deviation of the improved rate at the time of instability are calculated as and .

[0067] Step 7: Calculate the reliability index β in the time interval.

[0068] When the random fluctuations of the monitoring curve conform to the normal distribution, the value and the standard deviation of the deformation rate of the single measurement data in the selected time period can be calculated, and the specific formula is as follows:

[0069] , (5)

[0070] (6) wherein μ v , σv the improved deformation rate mean value and standard deviation in a certain equal time interval, v i the improved deformation rate of the adjacent two times of measurement in a certain equal time interval, the same parameters are used in the formula of the embodiment to represent the S-t curve and the T-t curve.

[0071] On this basis, the applicant statistically analyzes the improved tangent angle at the time of landslide failure, collects the monitoring data cases of landslide deformation instability that have occurred at home and abroad, and statistically analyzes the improved tangent angle at the time of instability according to the improved tangent angle method. The applicant has statistically analyzed 23 cases of landslide (slope) instability tangent angle that have occurred at home and abroad, and the specific results are shown in the graph as Figure 3 and Figure 4 Figure 3 The abscissa is the improved tangent angle at the time of instability of the monitoring curve of the landslide at the time of sliding according to the statistics of the landslides that have occurred, Figure 3 The ordinate is the number of landslides, Figure 4 The abscissa is the improved deformation rate at the time of instability according to the statistics of the landslides that have occurred, Figure 4 The ordinate is the number of landslides, according to the histogram, it can be approximately considered that the statistical data of the instability tangent angle of the landslide failure does not satisfy the normal distribution condition under normal circumstances, but the improved rate approximately satisfies the normal distribution characteristics, then the mean value and standard deviation of the improved rate at the time of instability can be calculated according to the statistical characteristics and .

[0072] In this step, the improved rate statistical data at the time of landslide instability is taken as the "resistance" of the slope instability, the rate statistical data in a certain fixed time period is taken as the "load effect" of the landslide instability, and when the improved deformation rate of the monitoring data and the improved deformation rate at the time of instability are subject to normal distribution, then according to the structure reliability theory, the reliability function of the landslide instability can be expressed as:

[0073] (7)

[0074] v and V respectively represent the improved deformation rate measured value in a certain equal time interval, V represents the improved rate value at the time of landslide instability, and v and V are both random variables subject to normal distribution.

[0075] The calculation formula of the reliability index β is:

[0076] (8) in the formula, the improved deformation rate mean value and standard deviation in a certain equal time interval, v ​Let represent the improved mean and standard deviation of deformation rates at the moment of landslide instability, based on case statistics. Z is the reliability function for landslide instability, specifically expressed as: v and V represent the improved deformation rate measurements within a certain equal time interval, and V represents the improved rate at the moment of landslide instability. Both v and V are random variables that follow a normal distribution.

[0077] Step 8: Calculate the failure probability based on the reliability index, and classify the landslide warning level according to the failure probability.

[0078] The formula for calculating the probability of damage in this step is:

[0079] (9),

[0080] In the formula, P f φ represents the probability of failure, φ represents the cumulative probability distribution function of the normal distribution, and β is the reliability index.

[0081] Table 1 shows the criteria for early warning of slope instability probability, as detailed below:

[0082] Table 1. Criteria for Early Warning of Slope Instability Probability

[0083] Early warning level Red early warning Orange early warning Orange early warning Blue early warning Criteria for discrimination P f ≥90%]]> 90% ≥ P f ≥ 60% 60% ≥ P f ≥ 30% P f <30%]]>

[0084] As shown in Table 1, in this step, a red alert is issued when the probability of failure is greater than or equal to 90%; an orange alert is issued when the probability of failure is greater than or equal to 60% and less than 90%; a yellow alert is issued when the probability of failure is greater than or equal to 30% and less than 60%; and a blue alert is issued when the probability of failure is less than 30%. The blue alert is the lowest level, and the red alert is the highest. This plan uses a tiered warning system for landslide failure probabilities, allowing for the development of different safety measures for different levels. This facilitates tiered management by relevant personnel and ensures the safety of life and property.

[0085] The following is an example of using the above method for landslide early warning:

[0086] Taking the deformation data of a large bedding rock landslide in the Three Gorges Reservoir area as an example, this landslide is currently in a continuous deformation stage. The deformation intensifies during the annual flood season as the water level drops, and then tends to stabilize after the flood season. GNSS monitoring equipment was installed on the surface of the landslide, with a data acquisition frequency of 1 hour per acquisition. To verify the applicability of this method, 2113 monitoring data points from May to July 2024 were selected. The time displacement results are as follows: Figure 5According to the above results, one day (24 data) is set as a time interval, the deformation rate of each monitoring data in each day is counted according to formula (1)-(3), the mean and standard deviation of the improvement rate in each equal time interval are calculated by using formula (5), (6), and the comparison results are as shown in the following table: Figure 6 As shown in the figure, the points with large failure probability are mostly located at the position of sharp increase in deformation, from May 24 to July 24, there are three yellow warnings, which is consistent with the actual monitoring results.

[0087] In the scheme, the randomness of the monitoring data is monitored to reflect the probability of landslide instability, the stability of the slope body is monitored by using the random fluctuation, the tedious data processing of the traditional method is avoided, and the reliability and accuracy of the early warning and prediction are higher.

[0088] The above is only an embodiment of the present application, and the specific technical solutions and / or common knowledge of the characteristics in the scheme are not described in detail. It should be pointed out that for those skilled in the art, without departing from the technical scheme of the present application, a number of modifications and improvements can be made. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances. The protection scope claimed in the present application should be subject to the content of its claims, and the specific embodiments and the like recorded in the specification can be used to explain the content of the claims.

Claims

1. A probability early warning method based on random fluctuations of landslide deformation monitoring curves, characterized in that, The application relates to a landslide deformation monitoring method. Step 1: obtaining landslide deformation monitoring data, and establishing a cumulative displacement-time curve, namely an S-t curve; Step 2: determining a uniform deformation stage of the S-t curve; and dividing the cumulative displacement-time curve into several equal time interval sections; Step 3: Calculate the average deformation rate during the uniform deformation phase and the deformation rate v of the two adjacent measurements within the same time interval i ; Step 4: Calculate the improved deformation rate v for the two adjacent measurements in the same time interval i / ; Step 5: improving the dimensions of the longitudinal and transverse coordinates of the S-t curve, and improving the original S-t curve into a T-t curve, wherein T represents the ratio of the current deformation rate to the rate of the uniform deformation stage; Step 6: If the improved deformation rate v i / is normally distributed, calculate the mean μ i / and standard deviation σ v of the deformation rates v v in the adjacent time intervals after improvement Step 7: calculating a reliability index beta in the equal time interval; Step 8: calculating a failure probability according to the reliability index, and dividing a landslide early warning level according to the failure probability; the calculation formula of the reliability index beta is as follows: where, are the mean and standard deviation of the improved deformation rate v i / are the mean and standard deviation of the improved deformation rate v i / at the time of landslide failure according to case statistics, and Z is the reliability performance function of landslide failure, which is specifically expressed as: , v and V are the mean and standard deviation of the improved deformation rate v i / at the time of landslide failure according to case statistics, i / , v and V are both random variables subject to normal distribution;​ The calculation formula of the failure probability is as follows: where P f denotes the failure probability, φ denotes the normal distribution cumulative probability distribution function, and β is the reliability index.

2. The probability early warning method based on landslide deformation monitoring curve random fluctuation according to claim 1, characterized in that: When the failure probability is greater than or equal to 90% in the step 8, red early warning is adopted; when the failure probability is greater than or equal to 60% and less than 90%, orange early warning is adopted; when the failure probability is greater than or equal to 30% and less than 60%, yellow early warning is adopted; When the failure probability is less than 30%, blue early warning is adopted; The blue early warning level is the lowest, and the red early warning level is the highest.

3. The probability early warning method based on random fluctuation of landslide deformation monitoring curve according to claim 1, characterized in that: In the step 2, the 3sigma criterion is adopted to determine the turning points of the uniform deformation stage and the accelerated deformation stage of the S-t curve.

4. The probability early warning method based on random fluctuation of landslide deformation monitoring curve according to claim 3, characterized in that: The three consecutive monitoring data in step 2 satisfy the following requirements, , the n-th monitoring data is considered as the starting point of the accelerated deformation stage, wherein: v n+1 , v n+2 , v n+3 respectively represent the three consecutive monitoring data after the n-th data point in the S-t curve, represents the mean of the rate calculated from the first n data points, represents the standard deviation of the rate calculated from the first n data points.

Images