Debris flow monitoring, positioning and early warning system based on infrasonic waves

Abstract

The invention discloses a debris flow monitoring, positioning and early warning system based on infrasonic waves. The debris flow monitoring, positioning and early warning system comprises the following steps that S1, infrasonic sensor arrays are arranged for debris flow prone areas in mountainous areas with complex construction periods; s2, transmitting data to a data processing center through 4G or Beidou wireless communication technology by collecting infrasound signals in real time; s3, carrying out data processing, analysis and identification on the collected infrasound signals; s4, on the basis of comprehensively considering an infrasound signal energy attenuation model and a time difference of arrival (TDOA) algorithm, calculating the position of a debris flow infrasonic wave source by utilizing energy attenuation data of infrasound signals received by a plurality of sensors and the time difference; s5, the system automatically triggers an early warning mechanism according to the monitoring result and a set threshold value to give an early warning to the public; and S6, displaying the debris flow infrasound signal waveform, the debris flow position and the early warning state in real time through a visual monitoring platform.

Description

Technical Field

[0001] The present invention belongs to the technical field of debris flow monitoring, and in particular relates to a debris flow monitoring, positioning and early warning system based on infrasound waves. Background Art

[0002] Debris flows often occur in mountainous areas with rugged terrain, such as mountain ridges and valleys. These flows, triggered by severe natural events like heavy rainfall and snowfall, are a unique mix of earth, rock, and gravel. Debris flows are sudden and destructive, often collapsing numerous structures and structures in residential areas, posing a significant threat. my country is also prone to frequent debris flows, primarily due to persistent heavy rainfall. Southwest my country, with its rich geological structure and mountainous terrain, is prone to intense crustal movement, and receives abundant rainfall during the rainy season. This makes debris flows more likely to occur and directly threaten the lives and property of local residents. Traditional debris flow monitoring methods rely primarily on meteorological and seismic methods, which suffer from slow response and low accuracy. Monitoring is particularly challenging in complex terrain and challenging environments. Accurately capturing and analyzing infrasound signals from debris flows in real time under these complex conditions, combined with comprehensive data processing and coordinated alarming, has become a pressing technical challenge in this field.

[0003] During the rapid movement of debris flow fluid, collisions between particles and between particles and fluid couple with the air, emitting low-frequency sound waves. The propagation speed of infrasound in air is much higher than that of debris flow motion, allowing for long propagation distances and minimal energy attenuation. It offers advantages such as a long warning lead time, lack of contact, and poor instrument integrity. Therefore, the use of infrasound signals to monitor and locate debris flows offers a new technical approach to addressing the shortcomings of existing technologies. As national infrastructure development continues to expand into inland mountainous areas, the challenge of safely and reliably building road interchanges in complex and dangerous mountainous areas presents significant challenges. Therefore, effective and rational monitoring, location, and early warning of debris flows are crucial for disaster prevention and mitigation efforts within mountainous infrastructure development. Summary of the Invention

[0004] The present invention provides a debris flow monitoring, positioning and early warning system based on infrasound waves, which aims to monitor and warn of debris flow disasters in mountainous areas to reduce the damage caused by the debris flow.

[0005] To this end, the present invention adopts the following technical solutions:

[0006] The debris flow monitoring, positioning and early warning system based on infrasound includes the following steps:

[0007] Step S1: deploying an infrasound sensor array in a debris flow prone area in a complex mountainous area during the construction period;

[0008] Step S2: The infrasound sensor collects infrasound signals in real time and transmits the data to the data processing center through the data transmission module;

[0009] Step S3: The data processing center processes and analyzes the collected infrasound signals, extracts the infrasound signals of debris flow from the infrasound signals, and determines the time and energy intensity characteristic information of the signals received at different stations when the debris flow event occurs;

[0010] Step S4: Based on a comprehensive consideration of the infrasound signal energy attenuation model and the Time Difference of Arrival (TDOA) algorithm, the location of the debris flow infrasound source is calculated using the energy attenuation data and time difference of the infrasound signals received by multiple infrasound sensors;

[0011] Step S5: The data processing center automatically triggers the early warning mechanism based on the monitoring results and the set thresholds to warn the public;

[0012] Step S6: The debris flow infrasound signal waveform, debris flow location and warning status are displayed in real time through the visual monitoring platform.

[0013] Furthermore, step S1 includes the following contents:

[0014] Step S11: deploy multiple infrasound sensor sites near debris flow-prone areas, each site is equipped with signal acquisition equipment, and each infrasound sensor is uniformly timed via GPS to ensure sensor time synchronization;

[0015] Step S12: The placement of infrasound sensors must comprehensively consider terrain characteristics, debris flow channels, and historical disaster distribution to ensure that the infrasound sensors can accurately capture the infrasound wave signals generated by the debris flow;

[0016] Step S13: The spacing between infrasound sensor sites is adjusted according to the scale and complexity of the terrain. To ensure the accuracy of positioning, the spacing between adjacent sensors is 1-5 km.

[0017] Furthermore, step S3 includes the following contents:

[0018] Step S31: using a Chebyshev II filter to perform high-pass and then low-pass filtering on the infrasound signal collected in step S2 to highlight the infrasound characteristics;

[0019] Step S32: Using the dB3 wavelet basis to perform noise reduction processing on the signal filtered in S31, decomposing the signal frequency band and removing random noise;

[0020] Step S33: extracting characteristic frequency bands of debris flow from the infrasound signal filtered and denoised in S32 using Fourier transform and time-frequency analysis;

[0021] Step S34: Use the ResNet50+SVM model to perform image recognition and classification on the time-frequency image of S33, automatically identifying the debris flow infrasound signal based on the deep-level characteristics of the infrasound signal, and deleting the data if the signal is not a debris flow signal;

[0022] Step S35: Outputting the signal time and energy intensity characteristic information received by different stations when the debris flow event occurs.

[0023] Furthermore, step S4 includes the following contents:

[0024] Step S41: Based on the infrasound signal energy attenuation model, the location of the debris flow infrasound source is calculated using the energy attenuation degree of the infrasound signals received by the multiple sensors extracted in step S3, wherein:

[0025] According to the principles of acoustics, the energy attenuation of infrasound waves can be expressed as:

[0026]

[0027] Where:

[0028]

[0029] Where E(d) is the energy (frequency) measured at a distance d, E0 is the initial energy at the source, n is usually 2, and this coefficient represents the geometric attenuation of energy with distance; α is the frequency-dependent attenuation coefficient, which depends on the medium properties and the frequency of the infrasound wave; f is the infrasound signal frequency; Q is the attenuation factor; and β is the shear wave velocity.

[0030] Step S42: Calculating the location of the debris flow infrasound source using the time difference between infrasound signals received by multiple infrasound sensors based on the Time Difference of Arrival (TDOA) algorithm;

[0031] Step S44: Output the coordinates of the debris flow source and the time of occurrence.

[0032] Furthermore, in step S42:

[0033] Select any infrasound sensor as the reference point R1, and calculate the time difference between other infrasound sensors and the reference point:

[0034] Δt i =t i -t1,i=2,3,…,n (1-3)

[0035] Where Δt i The infrasound signal reaches R i The time difference with R1;

[0036] The distance difference is calculated based on the time difference and the propagation speed of infrasound:

[0037] Δd i =v·Δt i (1-4)

[0038] Where Δd i Indicates the signal source (the location where the debris flow occurs) to R i The distance difference from R1;

[0039] For n receivers, according to the geometric relationship, the nonlinear equations are constructed as follows:

[0040] (1) Signal source (x, y) and the i-th sensor R i The distance between

[0041]

[0042] (2) The distance difference between the signal source and the i-th sensor and the first sensor is

[0043] Δd i =d i -d1,i=2,3,…,n (1-6)

[0044] (3) Substituting the distance difference into the positioning equation, we get:

[0045]

[0046] Step S43: Combine the energy attenuation models of multiple sites with the distance model and transform them into a mathematical optimization problem:

[0047]

[0048] (4) Using the Levenberg-Marquardt algorithm, the positioning equation is solved as follows:

[0049] 1) Construct the residual function R(x,y) of the positioning equation

[0050]

[0051] Among them, the first n-1 residuals correspond to the time positioning model, and the last n-1 residuals correspond to the energy decay model;

[0052] 2) Minimize the sum of squares of the residual function R(x,y) by iteratively updating the parameter P = (x,y). The steps are as follows:

[0053] ①Data initialization

[0054] At this stage, the parameter P is initialized (0) =(x (0) ,y (0)), set the damping factor λ value, the maximum number of iterations and the convergence threshold;

[0055] ②Iterative update

[0056] a. Calculate the residual vector R(P (k) )

[0057] b. Calculate the Jacobian matrix J(P (k) ), where each element is the partial derivative of the residual function with respect to x and y:

[0058]

[0059] Among them, the residual of the time difference positioning model is:

[0060]

[0061] The residual for the energy decay model is:

[0062]

[0063] c. Calculate the update step size Δp:

[0064] Δp=-(J T J+λI) -1 J T R (1-13)

[0065] Where I is the identity matrix;

[0066] d. Update parameters:

[0067] P (k+1) =P (k) +Δp (1-14)

[0068] e. Calculate the new residual sum of squares

[0069] S (k+1) =R(P (k+1) ) T R(P (k+1) ) (1-15)

[0070] f. Determine whether it converges:

[0071] If S (k+1) <S (k) Then reduce the damping factor λ; if S (k+1) >S (k) Then increase the damping factor λ;

[0072] g. Repeat the iteration until convergence is reached. The convergence condition is:

[0073] ||Δp||<ε (1-16)

[0074] ③ Output the calculation result F(x, y);

[0075] Step S44: Output the coordinates of the debris flow source and the time of occurrence.

[0076] The beneficial effects of the present invention are as follows: by deploying multiple infrasound sensors in mountainous areas, when a debris flow occurs, each infrasound sensor detects an infrasound signal. After noise reduction processing, the infrasound signal is distinguished as a debris flow signal, and the location of the debris flow is located based on the time difference between the signals received by different infrasound sensors, and then an early warning message is issued; the system operates fully automatically, without the need for human intervention, and is suitable for operation in remote areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 This is a diagram of a test device according to an embodiment of the present invention;

[0078] Figure 2 This is a schematic diagram of the overall module of the present invention;

[0079] Figure 3 It is a specific flow diagram of the present invention;

[0080] Figure 4 This is a time sequence diagram of the original infrasound signal of the debris flow test in this embodiment;

[0081] Figure 5 This is a graph showing the original infrasound signal of the debris flow test in this embodiment;

[0082] Figure 6 This is a spectrum diagram of the infrasound signal after filtering of the infrasound wave in the debris flow test in this embodiment;

[0083] Figure 7 This is a spectrum diagram of the infrasound signal after wavelet denoising of the debris flow test in this embodiment;

[0084] Figure 8 This is a time-frequency diagram of the infrasound signal after filtering and noise reduction in the debris flow test in this embodiment. DETAILED DESCRIPTION

[0085] The present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0086] The present invention proposes a debris flow monitoring, positioning and early warning system based on infrasound, which includes the following steps:

[0087] Step S1: deploying an infrasound sensor array in a complex mountainous area prone to debris flow.

[0088] Step S2: Infrasound signals are collected in real time and transmitted to a data processing center via 5G wireless communication technology.

[0089] Step S3: performing data processing, analysis and identification on the collected infrasound signals;

[0090] Step S4: Based on the Time Difference of Arrival (TDOA) algorithm, the location of the debris flow infrasound source is calculated using the time difference between the infrasound signals received by multiple sensors.

[0091] Step S5: The system automatically triggers the early warning mechanism based on the monitoring results and the set thresholds to warn the public.

[0092] Step S6: Through the visual monitoring platform, the debris flow infrasound signal waveform, debris flow location and warning status (such as Figure 4 and 5 ).

[0093] In this embodiment, step S1 includes the following contents:

[0094] Step S11: deploy multiple infrasound sensor sites near debris flow-prone areas, each site is equipped with necessary signal acquisition equipment, and each sensor is synchronized by GPS to ensure time synchronization;

[0095] Step S12: The placement of infrasound sensors must comprehensively consider terrain characteristics, debris flow channels, and historical disaster distribution to ensure that the sensors can accurately capture the infrasound wave signals generated by the debris flow;

[0096] Step S13: The spacing between sensor sites is adjusted according to the scale and complexity of the terrain. Usually, the spacing between sensors is 1-5 km to ensure the accuracy of positioning.

[0097] In this embodiment, step S3 includes the following contents:

[0098] Step S31: using a Chebyshev II filter to perform high-pass and then low-pass filtering on the infrasound signal collected in step S2 to highlight the infrasound characteristics;

[0099] Step S32: Use dB3 wavelet basis to perform noise reduction on the filtered signal to decompose the signal frequency band and remove random noise (such as Figure 6 and 7 );

[0100] Step S33: The filtered and denoised infrasound signal is processed by Fourier transform and time-frequency (e.g. Figure 8 ) Analyze and extract characteristic frequency bands of debris flow;

[0101] Step S34: Use the ResNet50+SVM model to perform image recognition and classification on the above time-frequency images. Automatically identify the debris flow infrasound signal based on the deep-level characteristics of the infrasound signal. The model performance is evaluated using accuracy, precision, recall, and F1. The performance evaluation results are as follows:

[0102] After the processed debris flow signal and environmental noise signal were extracted and classified by the ResNet50+SVM model, the performance evaluation indicators reached 88.60%, 92.19%, 89.39%, and 90.77% respectively;

[0103] Step S35: Outputting information such as the occurrence time and energy intensity characteristics of the debris flow event;

[0104] Furthermore, step S4 includes the following contents:

[0105] Step S41: Based on the infrasound signal energy attenuation model, the location of the debris flow infrasound source is calculated using the energy attenuation degree of the infrasound signals received by the multiple sensors extracted in step S3, wherein:

[0106] According to the principles of acoustics, the energy attenuation of infrasound can be expressed as:

[0107]

[0108] Where:

[0109]

[0110] Where E(d) is the energy (frequency) measured at a distance d, E0 is the initial energy at the source, n is usually 2, and this coefficient represents the geometric attenuation of energy with distance, α is the frequency-dependent attenuation coefficient, which depends on the medium properties and the frequency of the infrasound wave, f is the infrasound signal frequency, Q is the attenuation factor, and β is the shear wave velocity;

[0111] Step S42: Based on the Time Difference of Arrival (TDOA) algorithm, the location of the debris flow infrasound source is calculated using the time difference between the infrasound signals received by multiple sensors, where:

[0112] Select any sensor as the reference point R1, and calculate the time difference between other sensors and the reference point:

[0113] Δt i =t i -t1,i=2,3,…,n (1-3)

[0114] Where Δt i The infrasound signal reaches R i The time difference with R1;

[0115] The distance difference is calculated based on the time difference and the propagation speed of infrasound:

[0116] Δd i =v·Δt i (1-4)

[0117] Where Δd i Indicates the signal source (the location where the debris flow occurs) to R i The distance difference from R1;

[0118] For n receivers, according to the geometric relationship, the nonlinear equations are constructed as follows:

[0119] (1) Signal source (x, y) and the i-th sensor R i The distance between

[0120]

[0121] (2) The distance difference between the signal source and the i-th sensor and the first sensor is

[0122] Δd i =d i -d1,i=2,3,…,n (1-6)

[0123] (3) Substituting the distance difference into the positioning equation, we get:

[0124]

[0125] Step S43: Combine the energy attenuation models of multiple sites with the distance model and transform them into a mathematical optimization problem:

[0126]

[0127] (4) Using the Levenberg-Marquardt algorithm, the positioning equation is solved as follows:

[0128] 1) Construct the residual function R(x,y) of the positioning equation

[0129]

[0130] Among them, the first n-1 residuals correspond to the time positioning model, and the last n-1 residuals correspond to the energy decay model;

[0131] 2) Minimize the sum of squares of the residual function R(x,y) by iteratively updating the parameter P = (x,y). The steps are as follows:

[0132] ①Data initialization

[0133] At this stage, the parameter P is initialized (0) =(x (0),y (0) ), set the damping factor λ value, the maximum number of iterations and the convergence threshold;

[0134] ②Iterative update

[0135] a. Calculate the residual vector R(P (k) )

[0136] b. Calculate the Jacobian matrix J(P (k) ), where each element is the partial derivative of the residual function with respect to x and y:

[0137]

[0138] Among them, the residual of the time difference positioning model is:

[0139]

[0140] The residual for the energy decay model is:

[0141]

[0142] c. Calculate the update step size Δp:

[0143] Δp=-(J T J+λI) -1 J T R (1-13)

[0144] Where I is the identity matrix;

[0145] d. Update parameters:

[0146] P (k+1) =P (k) +Δp (1-14)

[0147] e. Calculate the new residual sum of squares

[0148] S (k+1) =R(P (k+1) ) T R(P (k+1) ) (1-15)

[0149] f. Determine whether it converges:

[0150] If S (k+1) <S (k) Then reduce the damping factor λ; if S (k+1) >S (k) Then increase the damping factor λ;

[0151] g. Repeat the iteration until convergence is reached. The convergence condition is:

[0152] ||Δp||<ε (1-16)

[0153] ③Output the calculation result F(x, y).

[0154] Step S44: Output the coordinates of the debris flow source, occurrence time, and propagation direction. The results are as follows:

[0155] Serial number Actual occurrence point (distance from sensor No. 1) Anchor Point Absolute error 1 (1m, 1m) (0.88m, 0.74m) (0.12m, 0.26m) 2 (1m, 1m) (0.92m, 0.83m) (0.08m, 0.17m) …… …… …… …… 43 (1m, 1m) (0.94m, 0.91m) (0.06m, 0.09m) 44 (1m, 1m) (0.78m, 0.81m) (0.28m, 0.19m)

[0156] The test results showed that all 44 tests successfully triggered alarms, with positioning errors within acceptable limits, demonstrating the system's feasibility and accuracy. The waveforms of the signal propagating along the route, as well as the test results of the positioning and monitoring system, indicate that the proposed debris flow disaster monitoring and early warning technology solution and system are feasible, with good detection and alarm performance.

Claims

1. The debris flow monitoring, positioning and early warning system based on infrasound is characterized by: The positioning warning system includes the following steps: Step S1: deploying an infrasound sensor array in a debris flow prone area in a complex mountainous area during the construction period; Step S2: The infrasound sensor collects infrasound signals in real time and transmits the data to the data processing center through the data transmission module; Step S3: The data processing center processes and analyzes the collected infrasound signals, extracts the infrasound signals of debris flow from the infrasound signals, and determines the time and energy intensity characteristic information of the signals received at different stations when the debris flow event occurs; Step S4: Based on a comprehensive consideration of the infrasound signal energy attenuation model and the Time Difference of Arrival (TDOA) algorithm, the location of the debris flow infrasound source is calculated using the energy attenuation data and time difference of the infrasound signals received by multiple infrasound sensors; Step S5: The data processing center automatically triggers the early warning mechanism based on the monitoring results and the set thresholds to warn the public; Step S6: The debris flow infrasound signal waveform, debris flow location and warning status are displayed in real time through the visual monitoring platform.

2. The debris flow monitoring, positioning and early warning system based on infrasound according to claim 1 is characterized in that: The step S1 includes the following contents: Step S11: deploy multiple infrasound sensor sites near debris flow-prone areas, each site is equipped with signal acquisition equipment, and each infrasound sensor is uniformly timed via GPS to ensure sensor time synchronization; Step S12: The placement of infrasound sensors must comprehensively consider terrain characteristics, debris flow channels, and historical disaster distribution to ensure that the infrasound sensors can accurately capture the infrasound wave signals generated by the debris flow; Step S13: The spacing between infrasound sensor sites is adjusted according to the scale and complexity of the terrain. To ensure the accuracy of positioning, the spacing between adjacent sensors is 1-5 km.

3. The debris flow monitoring, positioning and early warning system based on infrasound according to claim 1 is characterized in that: The step S3 includes the following contents: Step S31: using a Chebyshev II filter to perform high-pass and then low-pass filtering on the infrasound signal collected in step S2 to highlight the infrasound characteristics; Step S32: Using the dB3 wavelet basis to perform noise reduction processing on the signal filtered in S31, decomposing the signal frequency band and removing random noise; Step S33: extracting characteristic frequency bands of debris flow from the infrasound signal filtered and denoised in S32 by Fourier transform and time-frequency analysis; Step S34: Use the ResNet50+SVM model to perform image recognition and classification on the time-frequency image of S33, automatically identifying the debris flow infrasound signal based on the deep-level characteristics of the infrasound signal, and deleting the data if the signal is not a debris flow signal; Step S35: Outputting the signal time and energy intensity characteristic information received by different stations when the debris flow event occurs.

4. The debris flow monitoring, positioning and early warning system based on infrasound according to claim 1 is characterized in that: The step S4 includes the following contents: Step S41: Based on the infrasound signal energy attenuation model, the location of the debris flow infrasound source is calculated using the energy attenuation degree of the infrasound signals received by the multiple sensors extracted in step S3, wherein: According to the principles of acoustics, the energy attenuation of infrasound waves can be expressed as: Where: Where E(d) is the energy (frequency) measured at a distance d, E0 is the initial energy at the source, n is usually 2, and this coefficient represents the geometric attenuation of energy with distance; α is the frequency-dependent attenuation coefficient, which depends on the medium properties and the frequency of the infrasound wave; f is the infrasound signal frequency; Q is the attenuation factor; and β is the shear wave velocity. Step S42: Calculating the location of the debris flow infrasound source using the time difference between infrasound signals received by multiple infrasound sensors based on the Time Difference of Arrival (TDOA) algorithm; Step S44: Output the coordinates of the debris flow source and the time of occurrence.

5. The debris flow monitoring, positioning and early warning system based on infrasound according to claim 4 is characterized in that: In the step S42: Select any infrasound sensor as the reference point R1, and calculate the time difference between other infrasound sensors and the reference point: Δt i =t i -t1,i=2,3,…,n (1-3) Where Δt i The infrasound signal reaches R i The time difference with R1; The distance difference is calculated based on the time difference and the propagation speed of infrasound: Δd i =v·Δt i (1-4) Where Δd i Indicates the signal source (the location where the debris flow occurs) to R i The distance difference from R1; For n receivers, according to the geometric relationship, the nonlinear equations are constructed as follows: (1) Signal source (x, y) and the i-th sensor R i The distance between (2) The distance difference between the signal source and the i-th sensor and the first sensor is Δd i =d i -d1,i=2,3,…,n (1-6) (3) Substituting the distance difference into the positioning equation, we get: Step S43: Combine the energy attenuation models of multiple sites with the distance model and transform them into a mathematical optimization problem: (4) Using the Levenberg-Marquardt algorithm, the positioning equation is solved as follows: 1) Construct the residual function R(x,y) of the positioning equation Among them, the first n-1 residuals correspond to the time positioning model, and the last n-1 residuals correspond to the energy decay model; 2) Minimize the sum of squares of the residual function R(x,y) by iteratively updating the parameter P = (x,y). The steps are as follows: ①Data initialization At this stage, the parameter P is initialized (0) =(x (0) ,y (0) ), set the damping factor λ value, the maximum number of iterations and the convergence threshold; ②Iterative update a. Calculate the residual vector R(P (k) ) b. Calculate the Jacobian matrix J(P (k) ), where each element is the partial derivative of the residual function with respect to x and y: Among them, the residual of the time difference positioning model is: The residual for the energy decay model is: c. Calculate the update step size Δp: Δp=-(J T J+λI) -1 J T R (1-13) Where I is the identity matrix; d. Update parameters: P (k+1) =P (k) +Δp (1-14) e. Calculate the new residual sum of squares S (k+1) =R(P (k+1) ) T R(P (k+1) ) (1-15) f. Determine whether it converges: If S (k+1) <S (k) Then reduce the damping factor λ; if S (k+1) >S (k) Then increase the damping factor λ; g. Repeat the iteration until convergence is reached. The convergence condition is: ||Δp||<ε (1-16) ③ Output the calculation result F(x, y); Step S44: Output the coordinates of the debris flow source and the time of occurrence.

Images