A method for identifying irreversible displacement of landslide based on acceleration

Abstract

The present invention relates to the field of geological disaster monitoring technology and discloses a method for identifying irreversible displacement of landslides based on acceleration. The method comprises the following steps: Step 1: obtaining an abnormal acceleration data sequence a(n) and calculating the comprehensive waveform symmetry coefficient S of the abnormal acceleration data sequence; Step 2: performing discrete integration on the obtained abnormal acceleration data sequence to obtain discrete velocity integral results and final displacement integral results; Step 3: calculating the proportion of data with positive values ​​in the discrete velocity integral results obtained in Step 2; and Step 4: calculating a landslide irreversible displacement judgment index and comparing the obtained irreversible displacement judgment index with a preset threshold to determine whether the landslide has generated irreversible displacement. This method utilizes acceleration monitoring data to effectively identify spatial irreversible displacement generated by a landslide, greatly improving the accuracy and reliability of landslide early warning and forecasting using acceleration monitoring data.

Description

Technical Field

[0001] The present invention relates to the technical field of geological disaster monitoring, and in particular to an acceleration-based landslide irreversible displacement identification method. Background Art

[0002] Landslides, as a common geological disaster, pose a serious threat to human life and property. In recent years, with the rapid development of sensing technology, acceleration monitoring methods have been widely used in landslide deformation monitoring and early warning. Acceleration monitoring uses accelerometers to measure the acceleration changes generated by a landslide during deformation, thereby analyzing the deformation characteristics and stability of the landslide. Landslide deformation characteristics include spatially irreversible displacement characteristics.

[0003] Commonly used accelerometers include MEMS accelerometers, fiber optic accelerometers, and piezoelectric accelerometers. These sensors can capture acceleration changes in the landslide body in the X, Y, and Z directions, providing multidimensional data support for landslide deformation analysis. Accelerometers are highly precise and can detect minute acceleration changes (typically on the order of 0.1 to 1 mg). This allows them to capture weak signals in the early stages of landslide deformation. Accelerometers are also small and lightweight, making them easy to deploy in complex terrain. Compared to traditional displacement monitoring methods, acceleration monitoring allows 24-hour, all-weather monitoring, unaffected by factors such as terrain, vegetation obstruction, sunlight, and weather, making it easier to capture early warning signs of landslides. Although acceleration monitoring has unique technical advantages compared to other monitoring technologies such as GNSS and stress, it also has certain limitations and disadvantages. For example, it is susceptible to interference from environmental noise. Accelerometers are extremely sensitive to environmental vibrations. Traffic, human mechanical operations, strong winds, etc. may all generate interference signals. Especially in areas with frequent human activities, noise signals may mask the actual landslide deformation signals, resulting in an increased false alarm rate. Currently, noise can be suppressed through digital filtering and pattern recognition technology, but completely eliminating interference remains challenging. At the same time, acceleration sensors, especially MEMS devices, have zero drift problems. Their performance may gradually degrade during long-term monitoring. Harsh outdoor environments (temperature changes, humidity, corrosion, etc.) will accelerate this process, affecting the long-term reliability of monitoring data.

[0004] In addition, it is difficult to quantitatively analyze the motion state characteristics of landslides using acceleration monitoring data. The reason is that acceleration monitoring directly obtains acceleration values ​​rather than displacement. To convert them into more intuitive displacement or deformation in engineering, complex integration operations are required. The integration process will amplify measurement errors and noise effects, resulting in inaccurate displacement estimation results, affecting the accuracy and reliability of landslide warnings. At the same time, the acceleration change characteristics of landslides under different geological conditions vary significantly, making it difficult to establish a unified warning threshold and warning model. Currently, the threshold is mostly determined by relying on historical data and similar engineering experience. Summary of the Invention

[0005] The present invention aims to provide a method for identifying irreversible displacement of landslides based on acceleration, so as to solve the problems of high false alarm rate, low accuracy and low reliability in current landslide monitoring using acceleration.

[0006] To achieve the above-mentioned object, the present invention adopts the following technical solution: a method for identifying irreversible displacement of landslide based on acceleration, comprising:

[0007] Step 1: Obtain the abnormal acceleration data sequence a(n) and calculate the comprehensive waveform symmetry coefficient S of the abnormal acceleration data sequence;

[0008] Step 2: Perform discrete integration on the acquired abnormal acceleration data sequence to calculate the discrete velocity integral result And the final displacement integration result ;

[0009] Step 3: Calculate the value in step 2 The proportion of data with positive median values ;

[0010] Step 4: Calculate the irreversible displacement index of landslide and compare the obtained irreversible displacement judgment index with the preset threshold value to determine whether the landslide has produced irreversible displacement; the irreversible displacement judgment index The calculation formula is: , among which, the indicator The closer it is to 1, the higher the probability that the landslide will produce irreversible displacement.

[0011] Based on a large number of actual landslide acceleration monitoring cases, the applicant found that when a landslide does not become unstable, it often monitors a large amount of abnormal pulse vibration acceleration data when it is disturbed by the environment, which can lead to a large number of false alarms. At the same time, during the actual deformation, instability and destruction of a landslide, spatial irreversible displacement will inevitably be generated, and the generation of spatial irreversible displacement will inevitably be accompanied by abnormal acceleration data. In other words, abnormal acceleration data will be generated under both environmental interference and real irreversible displacement conditions. The abnormalities caused by environmental interference will affect the identification of the real irreversible displacement of the landslide. By constructing a landslide irreversible displacement judgment index, the applicant effectively reduces the impact of abnormal environmental pulse vibration signals, effectively identifies the real irreversible displacement generated by the landslide, and significantly improves the accuracy and reliability of landslide monitoring and early warning using acceleration.

[0012] Preferably, in step 1, the abnormal acceleration data sequence is obtained , construct the comprehensive waveform symmetry coefficient S, the calculation formula of S is:

[0013]

[0014] Where, is the weight of the positive and negative half-cycle energy ratio, which is used to determine whether the pulse waveform is symmetrical up and down. is the normalized energy ratio of positive and negative half cycles, Peak amplitude ratio weight, is the peak amplitude ratio, is the envelope symmetry weight, is the normalized standard envelope symmetry.

[0015] Preferably, in step 1 .

[0016] Preferably, the The calculation methods include:

[0017] Calculate the positive and negative half-cycle energy ratio , the calculation formula is:

[0018] Where, is the energy ratio of the positive and negative half cycles, is the abnormal acceleration data sequence; normalization processing is performed to obtain , The formula is: ;in, The closer the value is to 1, the higher the waveform symmetry is. The formula is: , represents the positive peak value, which represents the maximum positive value in a(n); Represents the negative peak value, representing the negative maximum value in a(n), that is, the absolute maximum value among all negative values; Values ​​closer to 1 indicate a more symmetrical signal.

[0019] Preferably, the The calculation method includes: extracting the waveform envelope value corresponding to a(n) ; Use the waveform envelope value E(n) to calculate the envelope symmetry of the corresponding waveform , , where N is the length of a(n) data sequence, the envelope symmetry Perform normalization processing to obtain the normalized standard envelope symmetry , the calculation formula of normalized standard envelope symmetry is: .

[0020] Preferably, after filtering, denoising and eliminating errors on the abnormal acceleration data sequence a(n) in step 2, discrete integration is performed to obtain the velocity change value. The calculation process is as follows:

[0021] Assume that the acceleration sampling frequency is f, and the acceleration time series after data processing is , perform discrete Fourier transform on the time series:

[0022] ; The calculated frequency domain results To perform integration:

[0023]

[0024] in:

[0025]

[0026] After obtaining the frequency domain integration results, the corresponding velocity time series can be directly obtained by using the inverse Fourier transform:

[0027] Due to the large mass and significant inertia of the landslide, the acceleration changes generated during the actual sliding process are mostly low-frequency signals. Filtering and noise reduction processing is first performed to prevent white noise from being amplified in the subsequent integration to calculate velocity and displacement, affecting the accuracy of the data.

[0028] Preferably, in step 2, the discrete velocity is integrated using a trapezoidal time domain integration method to obtain the final displacement value. The specific calculation method is as follows:

[0029] , where The initial displacement is usually taken as 0, is the sampling time interval.

[0030] Preferably, in step 2, the DC component in the signal is deducted by using a sliding window mean, and the specific correction calculation formula is as follows:

[0031] , where is the acceleration value after removing the DC component; is the original acquired abnormal acceleration data sequence; is the sliding window width, and M is the sliding window radius, which is calculated as 1 / 4 of the number of sampling points in the signal period. The inherent zero bias of the accelerometer introduces a false DC component, which, after integration, produces false displacement unrelated to actual motion. Furthermore, the integration operation amplifies low-frequency noise. Removing the DC component from the acceleration sequence prevents the coupling of DC and low-frequency noise, which can lead to divergence in the displacement calculation results and a significant deviation from the true value of the displacement curve.

[0032] Preferably, the least squares method is used in step 2 to fit the error coefficient 、 , and then use the following formula to eliminate the linear trend error,

[0033] , where This is the acceleration time series ultimately used for velocity and displacement integral calculations. Landslide deformation and displacement processes are long, and the acquired acceleration series are often quite long. This long acquisition period can easily lead to non-true trend errors. Therefore, the linear trend error term is eliminated from the acceleration series to ensure the accuracy of subsequent calculation results.

[0034] Advantages of this solution:

[0035] When a landslide is disturbed by the environment and produces irreversible displacement in real space, it will inevitably induce abnormal changes in slope acceleration. In this scheme, on the one hand, the waveform symmetry coefficient is calculated from the acceleration waveform (the symmetry of the abnormal acceleration waveform generated by environmental interference is good, and the comprehensive waveform symmetry coefficient S is large, such as Figure 2 ), to determine whether environmental interference pulse vibration has occurred, which can effectively identify and eliminate abnormal interference vibration signals in landslide acceleration monitoring, greatly improving the effectiveness and accuracy of monitoring data; on the other hand, it can also be combined with the waveform symmetry coefficient (the symmetry of the abnormal acceleration waveform generated by the real irreversible displacement is poor, and the comprehensive waveform symmetry coefficient S is very small, such as Figure 3 ), and use the acceleration integral to calculate the velocity and displacement, and perform displacement identification based on the unique physical characteristics of velocity and displacement to determine whether irreversible deformation displacement has occurred; this solution combines the above two parameter indicators to jointly determine whether the landslide has caused irreversible displacement. The accuracy of this comprehensive indicator can reach 90%.

[0036] This scheme uses acceleration monitoring data to effectively identify the spatial irreversible displacement caused by landslides, improving the accuracy and reliability of landslide early warning and forecasting using acceleration monitoring data; enriching and expanding the application prospects of acceleration monitoring methods in landslide geological disaster monitoring and early warning; in this scheme, the motion state characteristics of the landslide are judged based on acceleration, and based on this characteristic, a comprehensive judgment is made on whether a landslide is about to occur, which makes the early warning more accurate and timely. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 Schematic diagram of a flow chart of an embodiment of the present invention.

[0038] Figure 2 This is a characteristic waveform diagram of abnormal pulse vibration acceleration generated under environmental interference in the present invention.

[0039] Figure 3 This is a characteristic waveform diagram of abnormal acceleration caused by the real irreversible displacement of the landslide in the present invention.

[0040] Figure 4 This is the abnormal acceleration pulse vibration identification result under environmental interference in an embodiment of the present invention (calculation example).

[0041] Figure 5 This is a diagram of the acceleration change (raw data) along the X direction during a landslide displacement process in an embodiment of the present invention.

[0042] Figure 6 This is a diagram showing the acceleration changes (after filtering and noise reduction) along the X direction during the displacement process of a landslide in an embodiment of the present invention.

[0043] Figure 7 This is a diagram of the acceleration change (which can be used for integral calculation) generated along the X direction during the displacement process of a landslide in an embodiment of the present invention.

[0044] Figure 8 for Figure 7 Schematic diagram of the acceleration curve after velocity integration.

[0045] Figure 9 for Figure 7 Schematic diagram of the acceleration curve after displacement integration. DETAILED DESCRIPTION

[0046] The following is further described in detail through specific implementation methods:

[0047] Example:

[0048] A landslide irreversible displacement identification method based on acceleration, such as Figure 1 As shown, including:

[0049] Step 1: Obtain the abnormal acceleration data sequence a(n) and calculate the comprehensive waveform symmetry coefficient S of the abnormal acceleration data sequence.

[0050] Landslides will produce acceleration changes when they produce real deformation displacement and when they are affected by external human factors. The applicant found through a large number of landslide model tests and field real landslide acceleration monitoring data that there is a significant difference between the acceleration generated by the landslide during real displacement and the acceleration generated when it is disturbed by the external environment. For example, Figure 2 and Figure 3 As shown, when a landslide produces real continuous displacement motion, the generated acceleration waveform has poor symmetry. The deformation produced by the landslide in its natural state must be displaced in the main sliding direction, and this displacement is irreversible and can only increase gradually. Therefore, when the displacement is calculated by integrating the acceleration, it must be a positive value. However, when a particle on the slope is subjected to vibration caused by external environmental interference, it must be a relatively regular damped motion, with the peak amplitude exponentially decaying, that is, the acceleration shows periodic changes, and the waveform has typical symmetry, with positive and negative acceleration changes in a symmetrical state. The applicant uses the above characteristics to identify abnormal landslide vibrations and determine whether the landslide has produced irreversible displacement.

[0051] in, Figure 2-Figure 9 The horizontal axis is the index value of the collected abnormal acceleration data sequence. For example, the horizontal axis value is 56500~57200, which means that there are 57200-56500+1=701 acceleration data in this abnormal acceleration data sequence a(n). The sequence length N is 701. Starting from the first data, they are represented as a(1), a(2), a(3)...a(701) respectively. Figure 5 、 Figure 6 、 Figure 7 The acc_x parameter in the upper right corner represents the acceleration data in the X direction. Figure 2-Figure 9 The unit of acceleration is g, which means the acceleration due to gravity, g=9.8m / s 2 .

[0052] This solution uses the acceleration vibration wave morphology curve to identify vibrations and determine whether irreversible displacement has occurred. The specific method is as follows:

[0053] Find the abnormal acceleration data sequence a(n): Assume that an acceleration value a is detected at a certain moment. When it is determined that this value is greater than 3 times the measurement accuracy of the acceleration sensor, start recording acceleration time history data until the standard deviation of the acceleration data residual in the last 2 seconds is less than 2 times the relative mean error (precision) of the corresponding acceleration sensor measurement, indicating that the landslide has returned to a stable state under the influence of external excitation. At this time, stop recording acceleration data and save the acceleration values ​​recorded during this period together with the acceleration values ​​2 seconds before the initial recording moment to form the abnormal acceleration data sequence a(n) for subsequent analysis and calculation. Among them, the applicant found through a large number of indoor tests that the current landslide body is stable only when the calculated residual standard deviation is less than or equal to 2 times the relative mean error (precision) of the corresponding acceleration sensor measurement.

[0054] (1)

[0055] Where, It is the relative mean error of the acceleration sensor measurement, and this value is related to the measurement range of the sensor.

[0056] According to the characteristics of the pulse vibration signal, the acceleration periodic change curve has typical symmetry, that is, the maximum acceleration amplitude value will change periodically between positive and negative values.

[0057] In this solution, the abnormal acceleration data sequence is captured , the calculation formula of the comprehensive waveform symmetry coefficient S is constructed as follows:

[0058] (2)

[0059] Where, is the comprehensive waveform symmetry coefficient, is the weight of the positive and negative half-cycle energy ratio, is the normalized energy ratio of positive and negative half cycles, Peak amplitude ratio weight, is the peak amplitude ratio, is the envelope symmetry weight, is the normalized standard envelope symmetry.

[0060] The positive and negative half-cycle energy ratio is mainly used to determine whether the pulse waveform is symmetrical. For discrete acceleration data, the specific calculation formula is as follows:

[0061] (3)

[0062] Where, is the energy ratio of the positive and negative half cycles, For all The square sum of the acceleration values, that is, the positive half-cycle energy, For all The square sum of the acceleration values, that is, the negative half-cycle energy, and n is the index value corresponding to the acceleration data.

[0063] Since the result of the calculation of the positive and negative half-cycle energy ratio is a positive real number, it is not convenient for quantitative analysis to determine the symmetry of the waveform. In this solution, normalization is performed and quantized into a value between 0 and 1, where 0 indicates that the waveform is completely asymmetric and 1 indicates that the waveform is completely symmetric. The closer the value is to 1, the higher the waveform symmetry. The formula is as follows:

[0064] (4)

[0065] Where, Positive and negative half-cycle energy ratio.

[0066] Peak Amplitude Ratio The calculation formula is as follows:

[0067] (5)

[0068] Where, The peak amplitude ratio is an indicator to measure the symmetry of the signal or image. The closer the value is to 1, the more symmetrical the signal is. represents the positive peak value, which represents the positive peak value in a(n), that is, the maximum value among all positive values; The negative peak value represents the maximum negative value in a(n), that is, the absolute maximum value among all negative values; Values ​​closer to 1 indicate a more symmetrical signal.

[0069] Based on the abnormal acceleration data sequence a(n), the waveform envelope value corresponding to a(n) is extracted using Hilbert transform :

[0070] (6)

[0071] Among them, Hilbert transform is an important signal processing tool, which is mainly used to convert real signals into analytical signals to extract the instantaneous amplitude (envelope) of the signal.

[0072] Furthermore, the envelope symmetry of the corresponding waveform is calculated using the waveform envelope value E(n) :

[0073] (7)

[0074] Where N is the length of the a(n) data sequence.

[0075] In addition, since there may be many unpredictable noise signals in the actual landslide monitoring process, it is easy to cause local distortion of the envelope, thus misjudging it as asymmetry. At the same time, since the sensor is prone to baseline drift under long-term working conditions, it will also cause the overall envelope to shift. In this scheme, the calculated envelope symmetry is further corrected and normalized.

[0076] The calculation formula for correcting and normalizing the envelope symmetry is:

[0077] (8)

[0078] Based on acceleration monitoring data obtained from a large number of landslide tests, it was found that among the various symmetry judgment indicators, the weight of waveform envelope symmetry is generally higher, while the weights of the positive and negative half-cycle energy ratio and the peak amplitude ratio are relatively lower. In this scheme, the weight coefficient distribution ratio shown in the following formula is the optimal configuration for judging the symmetry of pulse signals, that is, the accuracy rate of symmetry judgment is the highest.

[0079]

[0080] Based on the applicant's extensive test statistics, the comprehensive waveform symmetry coefficient of the pulse-shaped acceleration generated by landslide rock and soil when subjected to forced damped vibration under external environmental interference is mostly between 0.7 and 1.0 (with a 95% confidence level). For strictly symmetrical waveforms, the calculated comprehensive waveform symmetry coefficient is generally greater than 0.9. The algorithm for identifying landslide vibration pulse signals proposed in this invention calculates the signal's comprehensive waveform symmetry coefficient. If it is greater than or equal to 0.7, it is determined that the landslide has generated a pulse vibration signal under external environmental interference, but has not caused irreversible deformation or displacement. This algorithm can effectively identify abnormal interference vibration signals in landslide acceleration monitoring, significantly improving the effectiveness and accuracy of monitoring data.

[0081] The following is an example of calculating the comprehensive waveform symmetry coefficient S:

[0082] Take the vibration pulse acceleration change generated during a landslide deformation test as follows: Figure 4 shown.

[0083] ① Obtain abnormal acceleration data sequence , through the acceleration value a detected at any moment, when it is judged that the value is greater than 3 times the measurement accuracy of the acceleration sensor (±0.0005g), the acceleration time history data will be recorded, and the acceleration data will be stopped when the residual standard deviation of the acceleration data in the most recent 2s time period is less than or equal to 2 times the relative mean error (accuracy) of the corresponding acceleration sensor. If there are less than 100 data on the left and right sides, they will be automatically filled. Finally, 14 abnormal acceleration data sequence pulse vibration fragments are detected, as shown in Figure 4shown.

[0084] ② Calculate the comprehensive waveform symmetry coefficient S of each pulse vibration segment signal.

[0085] 1. Calculate the energy ratio of positive and negative half cycles

[0086] According to the calculation method proposed in the text of the present invention, the normalized positive and negative half-cycle energy ratios of each vibration pulse signal are calculated as shown in the following table.

[0087] Table SEQ Table\* ARABIC 1: Statistics of normalized positive and negative half-cycle energy ratio calculation results

[0088] Abnormal vibration pulse signal segment number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Positive and negative half-cycle energy ratio 0.83 0.84 0.99 0.93 0.84 0.94 0.78 0.96 0.92 0.83 0.81 0.99 0.88 0.87

[0089] 2. Calculate the peak amplitude ratio

[0090] The peak amplitude ratio calculated according to the formula is shown in the following table.

[0091] Table SEQ Table\* ARABIC 2: Peak amplitude ratio calculation results

[0092] Abnormal vibration pulse signal segment number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Peak Amplitude Ratio 0.64 0.64 0.78 0.85 0.7 0.98 0.66 0.81 0.87 0.56 0.47 0.9 0.6 0.64

[0093] 3. Calculate the standard envelope symmetry

[0094] The standard envelope symmetry coefficient calculated according to the formula is shown in the following table.

[0095] Table SEQ Table\* ARABIC 3: Standard Envelope Symmetry Coefficient Calculation Table

[0096] Abnormal vibration pulse signal segment number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Normalized standard envelope symmetry 0.76 0.82 0.60 0.68 0.70 0.69 0.75 0.77 0.68 0.74 0.61 0.79 0.57 0.75

[0097] 4. Calculate the comprehensive waveform symmetry coefficient

[0098] Abnormal vibration pulse signal segment number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Positive and negative half-cycle energy ratio 0.83 0.84 0.99 0.93 0.84 0.94 0.78 0.96 0.92 0.83 0.81 0.99 0.88 0.87 Peak Amplitude Ratio 0.64 0.64 0.78 0.85 0.7 0.98 0.66 0.81 0.87 0.56 0.47 0.9 0.6 0.64 Normalized standard envelope symmetry 0.76 0.82 0.60 0.68 0.70 0.69 0.75 0.77 0.68 0.74 0.61 0.79 0.57 0.75 Comprehensive waveform symmetry coefficient 0.75 0.79 0.69 0.75 0.72 0.79 0.74 0.81 0.75 0.72 0.61 0.84 0.62 0.75 Actual situation Environmental interference Environmental interference Environmental interference Environmental interference Environmental interference Environmental interference Environmental interference Environmental interference Environmental interference Environmental interference No environmental interference Environmental interference No environmental interference Environmental interference

[0099] According to the analysis of the results in the above table, the comprehensive waveform symmetry coefficients of most vibration pulse signals generated by environmental interference are greater than 0.7. The comprehensive waveform symmetry coefficient S of the present invention can accurately identify whether it is a real environmental interference pulse vibration signal. For segment No. 3, the calculated result is 0.69, which is very close to 0.7, resulting in a misjudgment. The overall accuracy rate is as high as 91%.

[0100] Step 2: Perform discrete integration on the acquired acceleration time series to calculate the discrete velocity integral result And the final displacement integration result , and use the comprehensive waveform symmetry coefficient mentioned above to jointly determine and identify the true irreversible displacement of the landslide.

[0101] The specific steps are as follows:

[0102] (1) Filter and denoise the acquired acceleration time series

[0103] The acceleration signals obtained by MEMS sensors installed in the field for a long time often contain a large amount of white noise. According to experimental data, white noise is mostly high-frequency noise. Due to the large mass and significant inertia of landslides, the acceleration changes generated during the actual sliding process are mostly low-frequency signals. Therefore, to prevent the white noise from being amplified and affecting the data accuracy during the subsequent integration to calculate velocity and displacement, it is necessary to first filter and reduce the noise of the acquired acceleration time series. By comparing multiple filtering and noise reduction methods, a comprehensive analysis shows that the Butterworth low-pass filter is the best way to remove high-frequency noise for the acceleration signal characteristics generated during the actual sliding process of landslides in the field. When the low-pass frequency threshold is set to 50Hz, the effect of removing high-frequency noise is the best.

[0104] (2) Removing the DC component from the acceleration sequence

[0105] The inherent zero bias of the accelerometer will introduce a false DC component, which will produce a false displacement unrelated to the actual movement after integration. At the same time, the integration operation has an amplifying effect on low-frequency noise. The unremoved DC component and the coupling of low-frequency noise will cause the displacement calculation results to diverge, and the displacement curve will deviate significantly from the true value. Therefore, in order to prevent the generation of error terms that increase linearly with time during time domain integration, it is necessary to remove the DC component in the acceleration data series in advance. The specific calculation method is as follows:

[0106] Through a large amount of actual monitoring data, it is found that the acceleration signal generated by the landslide during the sliding process is often a non-stationary fluctuation signal. The signal's mean, variance, and power spectrum density will continue to change over time. For such situations, it is advisable to use the sliding window mean to deduct the DC component in the signal. The specific correction calculation formula is as follows:

[0107] (9)

[0108] Where, is the acceleration data sequence after removing the DC component; is the original acquired abnormal acceleration data sequence; is the sliding window width, M is the sliding window radius, M is 1 / 4 of the number of sampling points in the signal period, and i is the relative offset within the sliding window, which is used to traverse all data points in the window and calculate the local mean.

[0109] (3) Verify the DC component removal effect

[0110] The zero-frequency component amplitude is obtained by Fourier transform to verify the DC component elimination effect. The calculation formula is as follows: when the zero-frequency component amplitude is closer to 0, the elimination effect is better. In the present invention, the zero-frequency component amplitude after the DC component is removed is required to be less than 0.1.

[0111] (10)

[0112] Where j is the imaginary unit in the complex number.

[0113] (4) Eliminating linear trend errors

[0114] Since the deformation and displacement process of landslides is often very long, the whole process of sliding damage can occur in a few hours in a fast case and can take several years in a slow case. Therefore, the landslide acceleration time series collected in landslide monitoring is often long. When collected for a long time, it is easy to produce non-true trend errors. Therefore, it is also necessary to eliminate the linear trend error. The present invention adopts a polynomial fitting method and uses the least squares method to fit the error coefficient 、 , and then use the following formula to eliminate the linear trend error.

[0115] (11)

[0116] Where, It is the acceleration data sequence that is finally used for velocity and displacement integral calculation.

[0117] (5) Velocity integral calculation

[0118] The acceleration time series after filtering, noise reduction and error elimination is used to perform discrete integration to obtain the velocity change value. After comparing multiple methods, the conventional time domain integration method is simpler, but the cumulative error after integration is relatively large. The frequency domain integration method used in this invention is more suitable for the calculation of motion displacement of relatively slow deformation such as landslides. The specific calculation method is as follows:

[0119] Assume that the acceleration sampling frequency is f, and the acceleration time series after data processing is , perform discrete Fourier transform on the time series:

[0120] (12)

[0121] Further calculate the frequency domain results To perform integration:

[0122] (13)

[0123] in:

[0124] (14)

[0125] Where, is the acceleration data sampling frequency.

[0126] After obtaining the frequency domain integration results, the corresponding velocity time series can be directly obtained by using the inverse Fourier transform:

[0127] (15)

[0128] (6) Displacement integral calculation

[0129] Furthermore, the trapezoidal time domain integration method is used to integrate the discrete velocity to obtain the final displacement value. The specific calculation method is as follows:

[0130] (16)

[0131] Where, The initial displacement is usually taken as 0, is the sampling time interval. The final cumulative displacement value is obtained by formula 16 .

[0132] In this step, velocity and displacement are calculated by integrating acceleration, and displacement identification is performed using the unique physical characteristics of velocity and displacement.

[0133] Step 3: Calculate the value in step 2 The proportion of data with positive median values .

[0134] Step 4: Calculate the irreversible displacement index of landslide The obtained irreversible displacement judgment index is compared with the preset threshold to determine whether the landslide has produced irreversible displacement.

[0135] According to the deformation and failure mechanism of landslides, when a landslide is deformed and destroyed by external factors such as rainfall, top loading or toe excavation, it will produce deformation displacement in the main sliding direction. This deformation displacement will gradually increase along the main sliding direction over time and is irreversible. That is, it is impossible for a landslide to produce a 5mm displacement in the main sliding direction at a certain moment and then produce a 5mm reverse displacement in the opposite direction of the main sliding direction at the next moment. Periodic displacement changes are only possible when a certain particle in the landslide rock and soil mass is stimulated by external factors and produces abnormal vibration. Therefore, based on the irreversible characteristic of deformation and failure displacement of landslides, which is a typical characteristic of landslides, acceleration monitoring data can be used to identify whether a landslide has produced real displacement.

[0136] Based on acceleration monitoring data obtained from a large number of landslide tests, it is found that the acceleration changes generated by the real deformation and displacement of the landslide due to deformation and destruction are significantly different from the pulse damping vibration caused by environmental interference. This is mainly manifested in the poor periodicity of the acceleration changes and the obvious asymmetry in the waveform. If the acceleration waveform shows symmetry, the final displacement obtained by integration should be very close to 0, which means that when the rock and soil mass vibrates due to environmental interference, it will eventually gradually return to the vibration starting position under the action of damping, that is, no actual effective displacement is generated. For obviously asymmetric acceleration curves, the displacement obtained by integration calculation is often significantly different from 0, which means that the mass has not returned to the starting displacement and ultimately produces a relatively obvious real spatial displacement. This is consistent with the irreversible displacement characteristics generated during the landslide instability process. Therefore, based on the above theory, the comprehensive waveform symmetry coefficient S mentioned above can be used as one of the reference indicators for identifying and determining whether the landslide has generated real spatial displacement.

[0137] Through a large number of model tests and in-situ test statistics, the applicant found that when the landslide produces real spatial displacement, the calculated acceleration comprehensive waveform symmetry coefficient S is generally less than 0.7. Therefore, when the calculated value of the comprehensive waveform symmetry coefficient is less than or equal to 0.7, it can be preliminarily determined that the landslide has produced real spatial irreversible displacement. However, when encountering a vibration pulse signal with some sampling data omitted, the symmetry coefficient of its waveform will also be low, so there may occasionally be misjudgment. Therefore, it is necessary to introduce other more physically meaningful parameter indicators on the basis of the waveform symmetry coefficient for joint judgment to improve the accuracy of the judgment results.

[0138] Therefore, according to the landslide instability failure mechanism, when the overall anti-sliding force of the landslide is unable to resist the sliding force, it will produce irreversible displacement deformation along the sliding surface under the action of gravity. That is, once the landslide produces real displacement deformation, theoretically, the projection value of the displacement value in the sliding surface inclination direction must be positive, and the displacement must be in a continuous increasing trend, and will not appear in the process of increasing and then significantly decreasing. At the same time, the sliding velocity must also be all positive values ​​(an increasing or decreasing trend is allowed). Based on the above-mentioned landslide instability motion principle, the sliding velocity and sliding displacement can be obtained by integrating the discrete acceleration values, and the change characteristics of the above-mentioned sliding displacement and sliding velocity are used as restriction conditions to comprehensively judge whether the landslide produces irreversible unidirectional displacement.

[0139] In this scheme, the comprehensive waveform symmetry coefficient of the abnormal acceleration data sequence is first calculated , discrete velocity integration results And the final displacement integration result , further statistics show The proportion of data with positive median values , and then use the following calculation formula to comprehensively calculate the landslide irreversible displacement judgment index :

[0140]

[0141] In the formula, the indicator The closer it is to 1, the higher the probability that the landslide will produce irreversible displacement. When the index reaches 0.7, it can be judged that the landslide has caused irreversible displacement. The accuracy of judgment using this comprehensive index can reach 90%.

[0142] The following are specific indicators for judging irreversible displacement: Calculation example:

[0143] Select the acceleration signal generated during the real displacement of a landslide, such as Figure 5 shown.

[0144] ① After filtering and denoising the acquired acceleration time series, Figure 6 shown.

[0145] ② Remove the DC component and trend term in the acceleration sequence to obtain the displacement acceleration value that can be used for integral calculation, such as Figure 7 shown.

[0146] ③Integrate the acceleration with velocity and displacement. The integral result is as follows: Figure 8 shown.

[0147] From the integral results, it can be seen that in the final sliding instability destruction stage, the displacement of the landslide experienced the whole process of "starting-acceleration-deceleration-uniform speed". Among the velocity values ​​obtained by the acceleration integral solution, the positive velocity Accounting for about 94%, the final displacement integral result >0, further using the vibration pulse waveform symmetry calculation method, the comprehensive waveform symmetry coefficient S of the acceleration waveform caused by the displacement change is calculated, and the positive and negative half-cycle energy ratio is calculated. is 0.90, the peak amplitude ratio The normalized standard envelope symmetry is 0.82. It is 0.132, and the comprehensive waveform symmetry coefficient S=0.90*0.15+0.82*0.35+0.132*0.65=0.51, which is much smaller than 0.7, indicating that the displacement acceleration waveform has poor symmetry and does not have the characteristics of pulse vibration.

[0148] Further calculate the landslide irreversible displacement judgment index :

[0149]

[0150] According to the judgment index, the judgment index of irreversible displacement of landslide If it is greater than 0.7, it can be determined that the landslide has produced irreversible displacement, which is consistent with the actual situation.

[0151] This scheme uses acceleration as monitoring data to effectively identify the spatial irreversible displacement caused by landslides, and uses acceleration monitoring data to distinguish the characteristics of irreversible displacement caused by landslides, which will greatly improve the accuracy and reliability of landslide early warning and forecasting using acceleration monitoring data, and will greatly enrich and expand the application prospects of acceleration monitoring methods in landslide geological disaster monitoring and early warning; this scheme judges the movement state characteristics of the landslide based on acceleration, and then makes a comprehensive judgment on whether a landslide is about to occur based on this characteristic, which makes the early warning more accurate and timely; this scheme judges the occurrence of irreversible displacement through acceleration, and then combines other indicators to make landslide early warning and forecast, avoiding the complex process of traditionally using acceleration to calculate landslide displacement, with less calculation amount and less computing resources occupied.

[0152] The above is only an embodiment of the present invention, and the common knowledge such as the specific technical solutions and / or characteristics in the solution are not described in detail here. It should be pointed out that for those skilled in the art, several variations and improvements can be made without departing from the technical solution of the present invention. In the present invention, unless otherwise clearly specified and limited, the terms "install", "connect", "connect", "fix" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected, or indirectly connected through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. The scope of protection claimed by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A method for identifying irreversible displacement of landslide based on acceleration, characterized in that: include: Step 1: Obtain the abnormal acceleration value data sequence, represented by a(n), where n is the index value corresponding to a certain acceleration value in the data sequence, and calculate the comprehensive waveform symmetry coefficient S of the abnormal acceleration data sequence; Step 2: Perform discrete integration on the acquired abnormal acceleration data sequence to calculate the discrete velocity integral result And the final displacement integration result ; Step 3: Calculate the value in step 2 The proportion of data with positive median values ; Step 4: Calculate the irreversible displacement index of landslide and compare the obtained irreversible displacement judgment index with the preset threshold value to determine whether the landslide has produced irreversible displacement; the irreversible displacement judgment index The calculation formula is: , among which, the indicator The closer it is to 1, the higher the probability that the landslide will produce irreversible displacement; In step 1, the abnormal acceleration data sequence is obtained , construct the comprehensive waveform symmetry coefficient S, the calculation formula of S is: Where, is the weight of the positive and negative half-cycle energy ratio, which is used to determine whether the pulse waveform is symmetrical up and down. is the normalized energy ratio of positive and negative half cycles, Peak amplitude ratio weight, is the peak amplitude ratio, is the envelope symmetry weight, is the normalized standard envelope symmetry.

2. The method for identifying irreversible displacement of landslide based on acceleration according to claim 1, characterized in that: In step 1 .

3. The method for identifying irreversible displacement of landslide based on acceleration according to claim 1, characterized in that: described The calculation methods include: Calculate the positive and negative half-cycle energy ratio , the calculation formula is: Where, is the energy ratio of the positive and negative half cycles, is the abnormal acceleration data sequence; normalization processing is performed to obtain , The formula is: ;in, The closer the value is to 1, the more symmetrical the waveform is.

4. The acceleration-based landslide irreversible displacement identification method according to claim 1, characterized in that: described The formula is: , represents the positive peak value, which represents the maximum positive value in a(n); The negative peak value represents the maximum negative value in a(n), that is, the absolute maximum value among all negative values; Values ​​closer to 1 indicate a more symmetrical signal.

5. The method for identifying irreversible displacement of landslide based on acceleration according to claim 1, characterized in that: described The calculation method includes: extracting the waveform envelope value corresponding to a(n) ; Use the waveform envelope value E(n) to calculate the envelope symmetry of the corresponding waveform , , where N is the length of a(n) data sequence, the envelope symmetry Perform normalization processing to obtain the normalized standard envelope symmetry , the calculation formula of normalized standard envelope symmetry is: .

6. The acceleration-based landslide irreversible displacement identification method according to claim 1, characterized in that: In step 2, after filtering, denoising and eliminating errors in the abnormal acceleration data sequence a(n), discrete integration is performed to obtain the velocity change value. The calculation process is as follows: Assume that the acceleration sampling frequency is f, and the acceleration time series after data processing is , perform discrete Fourier transform on this time series: ; The calculated frequency domain results To perform integration: in: After obtaining the frequency domain integration results, the corresponding velocity time series can be directly obtained by using the inverse Fourier transform: 。 7. The acceleration-based landslide irreversible displacement identification method according to claim 6, characterized in that: In step 2, the discrete velocity is integrated using the trapezoidal time domain integration method to obtain the final displacement value. The specific calculation method is as follows: , where Take 0 as the initial displacement, is the sampling time interval.

8. The acceleration-based landslide irreversible displacement identification method according to claim 6, characterized in that: In step 2, the DC component in the signal is deducted by using the sliding window mean value. The specific correction calculation formula is as follows: , where is the acceleration value after removing the DC component; is the original acquired abnormal acceleration data sequence; is the sliding window width, M is the sliding window radius, and M is 1 / 4 of the number of sampling points in the signal period.

9. The acceleration-based landslide irreversible displacement identification method according to claim 6, characterized in that: In step 2, the least square method is used to fit the error coefficient 、 , and then use the following formula to eliminate the linear trend error, , where is the acceleration time series that is ultimately used for velocity and displacement integral calculations.

Images