Method for monitoring settlement of coastal protection dams

Abstract

The application discloses a kind of coast protection embankment settlement monitoring methods, belong to hydraulic engineering safety monitoring technical field.The method includes: setting longitudinal monitoring section on the slope of embankment back sea side, arranging three measuring line fiber grating sensor array, and periodically collecting vertical displacement and horizontal inclination data by synchronous trigger unit;Utilize multi-source data fusion module to establish vertical displacement-horizontal inclination joint analysis model, and generate settlement distribution curve by Kalman filtering algorithm to eliminate tide interference;Based on hierarchical early warning rule triggers three to one level early warning signal, and is transmitted to terminal equipment.The application solves the problem that traditional monitoring method cannot effectively separate tidal interference, real-time is poor and multi-dimensional data collaborative analysis is insufficient, through multi-measuring line sensor deployment, dynamic filtering optimization and hierarchical early warning mechanism, realizes high-precision settlement monitoring, complex environment anti-interference and rapid emergency response, significantly improves the reliability and timeliness of dam safety monitoring.

Description

Technical Field

[0001] This invention relates to the field of water conservancy project safety monitoring technology, and in particular to a method for monitoring the settlement of coastal protection dikes. Background Technology

[0002] Coastal protection dikes are crucial facilities for resisting tidal erosion and storm surges, and their structural stability is directly related to the safety of life and property of people in coastal areas. Due to the long-term influence of complex environments such as tidal cycle loads, soil seepage, and wave impact, dikes are prone to problems such as uneven settlement, foundation erosion, and local instability.

[0003] Traditional settlement monitoring mainly relies on manual inspections, single-point settlement pile measurements, or fixed sensor networks, which have the following limitations:

[0004] Insufficient data timeliness: Manual inspections have long cycles (usually once a month), making it difficult to capture sudden settlements in a timely manner; although single-point sensors can monitor continuously, they lack the ability to conduct collaborative analysis across multiple measurement lines, and cannot fully reflect the overall deformation trend of the dam.

[0005] Environmental interference is difficult to separate: tidal periodic water level changes cause elastic deformation of dams, which, when superimposed with actual settlement, can easily lead to misjudgments. Existing technologies mostly use static filtering or simple thresholding methods, which are difficult to dynamically eliminate interference factors such as tide level and temperature.

[0006] The early warning mechanism is too simple: traditional methods usually trigger early warnings based on a single parameter (such as displacement), lacking coordinated judgment of multi-dimensional indicators such as settlement gradient and curvature change, resulting in a high rate of missed or false alarms.

[0007] Limited communication reliability: Coastal areas are often affected by severe environments such as typhoons and salt spray, making wired transmission prone to interruption. Wireless communication (such as GPRS) has insufficient bandwidth and stability, making it difficult to guarantee the real-time transmission of critical early warning information.

[0008] Existing improvement methods, such as using fiber optic grating sensors to enhance monitoring accuracy, are mostly deployed along a single line on the top of the embankment, making it difficult to capture stratified settlement characteristics. Some studies have attempted to introduce Kalman filtering algorithms, but without combining multi-source data such as tide level and tilt angle for dynamic parameter adjustment, the filtering effect is limited. Therefore, there is an urgent need for a embankment settlement monitoring method that can integrate multi-dimensional data, dynamically resist interference, and adapt to complex communication environments. Summary of the Invention

[0009] This invention overcomes the problems of traditional monitoring methods, such as ineffective separation of tidal interference, poor real-time performance, and insufficient multi-dimensional data collaborative analysis. Through the deployment of multi-line sensors, dynamic filtering optimization, and hierarchical early warning mechanism, it achieves high-precision settlement monitoring, anti-interference in complex environments, and rapid emergency response, significantly improving the reliability and timeliness of dam safety monitoring.

[0010] To achieve the above objectives, the present invention adopts the following solution:

[0011] A method for monitoring settlement of coastal protection dikes, comprising the following steps:

[0012] S1: A longitudinal monitoring section is set on the leeward slope of the main body of the dike. The longitudinal monitoring section extends along the dike axis. The longitudinal monitoring section includes a top survey line arranged on the top of the dike, a toe survey line arranged on the toe of the dike, and a middle survey line arranged on the middle platform of the dike body. Fiber Bragg grating sensor arrays are set at intervals on the top survey line, the toe survey line and the middle survey line. The fiber Bragg grating sensor arrays include multiple sensor nodes arranged at intervals along the dike axis. Each sensor node includes a displacement meter arranged in the vertical direction and an inclination sensor arranged in the horizontal direction.

[0013] S2: A data acquisition module is set up in the monitoring and control station. The data acquisition module is connected to three sets of fiber optic sensor arrays of the longitudinal monitoring section through armored optical cables. The data acquisition module includes a synchronization trigger unit and a data buffer unit. The synchronization trigger unit periodically sends synchronization acquisition commands to all sensor nodes. The data buffer unit receives and stores the vertical displacement data measured by the displacement gauge and the horizontal tilt data measured by the tilt sensor.

[0014] S3: Configure a multi-source data fusion module in the data processing center. The multi-source data fusion module receives vertical displacement data and horizontal tilt angle data transmitted by the data acquisition module, establishes a joint analysis model of vertical displacement and horizontal tilt angle, and uses the Kalman filter algorithm to eliminate periodic displacement interference caused by tidal changes and generate a settlement distribution curve in the dam axis direction.

[0015] S4: Deploy a settlement trend prediction module on the early warning server. The settlement trend prediction module is connected to the multi-source data fusion module and obtains the settlement distribution curve. When the vertical displacement data of any sensor node changes for N consecutive cycles, the change amount exceeds the change threshold D, a level 3 early warning signal is triggered. When the vertical displacement change gradient of three adjacent sensor nodes exceeds the gradient threshold B, a level 2 early warning signal is triggered. When the radius of curvature of the settlement distribution curve is less than the design allowable value C, a level 1 early warning signal is triggered.

[0016] S5: The early warning server sends early warning signals to terminal devices through communication base stations. The communication base stations adopt a dual-channel transmission mode of 4G network and Beidou short message. The terminal devices include computer terminals of the dam management center and mobile terminals of on-site maintenance personnel.

[0017] Preferably, step S3, establishing the joint analysis model of vertical displacement and horizontal tilt angle, includes the following steps:

[0018] The vertical displacement data is divided into discrete segments with 10-minute intervals according to the time series. The standard deviation of the displacement change rate and the slope of the linear fitting are calculated for each discrete segment to generate a sequence of vertical displacement characteristic parameters.

[0019] The horizontal tilt angle data is subjected to fast Fourier transform, and the energy ratio of the 0.1-0.5Hz frequency band is extracted as the dynamic feature of the tilt angle. When the energy ratio exceeds 60%, it is determined to be abnormal vibration interference.

[0020] Establish a correlation coefficient matrix between the vertical displacement characteristic parameter sequence and the tilt angle dynamic characteristic quantity. When the absolute value of the correlation coefficient is greater than 0.7, mark the vertical displacement data of the corresponding time period as the segment affected by tilt angle coupling.

[0021] In the Kalman filter algorithm, the measured value of the elastic modulus of the dam's main structure is introduced. The process noise covariance matrix of the filter is adjusted based on the product of the measured elastic modulus and the dynamic characteristic of the tilt angle. The adjustment formula is as follows: Where Q is the adjusted process noise covariance matrix, Q0 is the original process noise covariance matrix, E is the measured value of elastic modulus, and A is the dynamic characteristic of tilt angle;

[0022] The vertical displacement data of the section marked as being affected by tilt coupling is input into the adjusted Kalman filter to generate a settlement distribution curve that includes the influence of structural stiffness, while removing data points corresponding to abnormal vibration interference.

[0023] Preferably, in step S3, after generating the settlement distribution curve that includes the influence of structural stiffness, the following steps are also performed:

[0024] Based on the time periods corresponding to correlation coefficients with absolute values ​​greater than 0.7 in the correlation coefficient matrix, the vertical displacement characteristic parameter sequences of the top, middle, and toe survey lines of the embankment are extracted for the same time period, and the displacement compatibility index η among the three survey lines is calculated. , where σ top σ mid σ foot These are the standard deviations of the vertical displacement rates of the survey lines at the top, middle, and toes of the embankment, respectively.

[0025] The displacement compatibility index η is nonlinearly fitted with the measured value of the elastic modulus to establish an η-E relationship surface. When the curvature of the η-E relationship surface exceeds a preset threshold, it is determined that the main body of the dam has a tendency of layered settlement.

[0026] A stiffness anomaly detection unit is added to the settlement trend prediction module. The stiffness anomaly detection unit receives η-E relationship surface data. When a layered settlement trend is detected, a correction amount ΔR=γ×(η-η0) is automatically added to the curvature radius value of the corresponding embankment segment in the settlement distribution curve, where η0 is the benchmark coordination index during embankment construction and acceptance, and γ is the reciprocal of the width of the main cross-section of the embankment.

[0027] In step S4, the following judgment rule is added to the conditions for triggering the first-level warning signal: when the corrected radius of curvature value is less than the design allowable value and the curvature of the η-E relationship surface exceeds the preset threshold, the warning signal will be upgraded to the special warning signal.

[0028] When generating the settlement distribution curve, the multi-source data fusion module simultaneously outputs a three-dimensional deformation cloud map containing stratified settlement trend indicators. The distribution range of the η value is distinguished by color gradient in the three-dimensional deformation cloud map.

[0029] Preferably, in step S3, after generating the three-dimensional deformation cloud map containing the stratified settlement trend indicators, the following steps are also performed:

[0030] Among the top survey line, middle survey line and toe survey line of the dike, select the section where the sensor node with the lowest η value is located. Use a mobile detection vehicle to carry a ground-penetrating radar to scan the profile along the section. The ground-penetrating radar uses an 800MHz antenna to collect dielectric constant data at 0.2-meter intervals. At the same time, use a static cone penetration test device to measure the cone tip resistance value at a distance of 1 meter from the sensor node.

[0031] The dielectric constant data and cone tip resistance value are input into the multi-source data fusion module to establish a joint verification model of dielectric constant-cone tip resistance. The model outputs a correction coefficient μ for the compaction of the main soil mass of the dam. , where ε r q is a relative value of the dielectric constant. c This represents the cone tip resistance value.

[0032] The calculation formula of the displacement compatibility index η is adjusted according to the μ value, and is changed to η'=η×(1+0.2μ). The adjusted displacement compatibility index η' is then substituted back into the η-E relationship surface to determine the stratified settlement trend.

[0033] In step S2, the synchronization triggering unit of the data acquisition module dynamically adjusts the acquisition cycle according to the adjusted η' value: when η' < 0.6, the acquisition cycle is shortened to 30 minutes; when 0.6 ≤ η' ≤ 0.8, the cycle is maintained at 1 hour; when η' > 0.8, the acquisition cycle is extended to 2 hours.

[0034] In step S5, soil compaction correction information is added when the warning signal is sent. When μ < 1.2, a red identification code is added to the warning signal; when 1.2 ≤ μ ≤ 1.5, a yellow identification code is added; and when μ > 1.5, a green identification code is added.

[0035] The BeiDou short message transmission mode of the communication base station adjusts the transmission frequency according to the μ value: the red identification code corresponds to 3 times per hour, the yellow identification code corresponds to 1 time per hour, and the green identification code corresponds to 1 time every 3 hours.

[0036] Preferably, in step S3, the vertical displacement-horizontal tilt angle joint analysis model further includes the following processing steps:

[0037] The data processing center receives real-time tide data transmitted from the tide monitoring station. The tide monitoring station is located on the seaward side of the main body of the dike and 10 meters away from the toe of the dike. The tide monitoring station includes a pressure level gauge and a Beidou positioning module. The pressure level gauge measures the tide elevation in 5-minute intervals, and the Beidou positioning module records the latitude and longitude coordinates of the tide monitoring station.

[0038] The multi-source data fusion module timestamps and matches the tide level elevation with the vertical displacement data. It then uses a Kalman filter algorithm to extract the periodic displacement component caused by the tide level change. The extraction frequency of the periodic displacement component is 0.5-2 times / hour, and this component is removed from the vertical displacement data.

[0039] Based on the correlation analysis results between horizontal dip angle data and tidal level elevation, the weighting coefficients of the dip angle sensors in the joint analysis model are dynamically adjusted. The adjustment formula for the weighting coefficient α is as follows: , where ΔH is the difference in elevation between two adjacent tides, and k is the proportional factor corresponding to the permeability coefficient of the main soil of the dam.

[0040] In step S4, when the settlement trend prediction module triggers the secondary early warning signal, it simultaneously performs the following operations:

[0041] Extract the tide phase data of the tide monitoring station in the 24 hours before the current time, calculate the difference between the actual tide phase and the historical tide phase for the same period, and upgrade the level II warning signal to the level I warning signal when the difference exceeds 30 degrees and the vertical displacement change gradient exceeds 2 mm / m.

[0042] After generating the settlement distribution curve, based on the time delay correlation between the tide level data and the vertical displacement data, the displacement compensation of the sensor nodes within 200 meters of the tide level monitoring station in the settlement distribution curve is corrected. The displacement compensation is as follows: H inst H represents the current tide level elevation. meanβ is the average tide level for the month, and β is the tangent of the inclination angle of the seaward slope of the main body of the dike.

[0043] Preferably, in step S1, the sensor nodes of the fiber Bragg grating sensor array further include the following structural features:

[0044] A temperature compensation unit is set between the displacement gauge and the tilt sensor at each sensor node. The temperature compensation unit includes a platinum resistance temperature sensor that is in close contact with the surface of the dam body. The platinum resistance temperature sensor measures the contact temperature at the sensor node with a resolution of 0.1℃.

[0045] The output of the temperature compensation unit is connected to the data acquisition module, and the data buffer unit synchronously stores the temperature measurement data. The temperature measurement data has the same timestamp as the vertical displacement data and the horizontal tilt angle data.

[0046] In step S3, when establishing the vertical displacement-horizontal tilt angle joint analysis model, the multi-source data fusion module inputs temperature measurement data into the displacement compensation model. The displacement compensation model corrects the vertical displacement data in real time based on the temperature-displacement relationship curve. The formula for calculating the vertical displacement correction is as follows: , where α T T is the temperature sensitivity coefficient of the displacement gauge. inst The current temperature measurement value, T ref This is the reference temperature value during installation and commissioning;

[0047] Before processing the vertical displacement data, the Kalman filter algorithm first applies the corrected ΔD... c Accumulated amounts are deducted, and the deduction period is consistent with the 1-hour period of the synchronous acquisition command.

[0048] Preferably, in step S4, the early warning judgment of the settlement trend prediction module also includes the following analysis process:

[0049] Extract the current temperature measurement value T recorded by the temperature compensation unit. inst Compared with the reference temperature value T ref The temperature anomaly correction mode is activated when the absolute value of ΔT exceeds 5°C for three consecutive cycles.

[0050] In the temperature anomaly correction mode, a temperature compensation verification rule is added to the conditions for triggering the level 3 warning signal: when the vertical displacement data changes by more than 5 mm for three consecutive cycles, ΔD is calculated synchronously. c Compared with the measured displacement change ΔD measured ratio If the ratio K > 0.3, the warning signal will be downgraded to an observation-level alarm.

[0051] When determining whether the vertical displacement gradient of three adjacent sensor nodes exceeds 2 mm / m, the standard deviation σ of ΔT at these three nodes is calculated simultaneously. T When σ T When the temperature is >2℃, the gradient threshold is adjusted from 2 mm / m to 2.5 mm / m;

[0052] When calculating the radius of curvature of the settlement distribution curve, a cumulative temperature influence factor is introduced. , where ΔT i This represents the hourly ΔT value within the previous 24 hours of the current calculation period, where n is the number of samples. The design allowable value is multiplied by a correction factor [1 + 0.02 × (C)]. T -10)];

[0053] In step S5, the warning signal is accompanied by a temperature-related identifier code. When the absolute value of ΔT exceeds 10℃, a temperature anomaly marker is embedded in the warning signal. This marker triggers the mobile terminal of the terminal device to start the temperature calibration mode. The temperature calibration mode forcibly calls the ground temperature monitoring data of the most recent 3 hours and cross-verifies it with the data of the platinum resistance temperature sensor.

[0054] Preferably, in step S2, the synchronization triggering unit further includes a dynamic acquisition mode switching function, which operates as follows:

[0055] The data acquisition module continuously monitors the rate of change of the vertical displacement data. When the vertical displacement change of any sensor node exceeds 3 mm for two consecutive cycles, it automatically switches to emergency acquisition mode.

[0056] In emergency data acquisition mode, the synchronous triggering unit shortens the acquisition cycle to 10 minutes and prioritizes the transmission of data from all nodes within a 50-meter range upstream and downstream of the sensor node.

[0057] In emergency acquisition mode, the data caching unit activates a block storage mechanism, marking priority data as high-priority data packets. High-priority data packets are transmitted in real time through the 4G network channel, while the remaining data is stored in the local cache.

[0058] In step S4, when the settlement trend prediction module receives a high-priority data packet, it starts a fast prediction algorithm. The fast prediction algorithm uses the exponential smoothing method to extrapolate the data of the most recent 6 periods. When the vertical displacement acceleration of the extrapolation result exceeds 0.05 mm / min², a secondary warning signal is forcibly triggered.

[0059] The BeiDou short message transmission mode of the communication base station is switched to continuous transmission mode in emergency collection mode, and the continuous transmission interval is set to 5 minutes.

[0060] Preferably, in step S5, the warning signal processing of the terminal device further includes the following linkage control logic:

[0061] When the computer terminal of the dam management center receives a Level 1 early warning signal, it automatically activates the emergency control interface. This interface connects to the PLC controller of the dam's drainage gate, sending a gate opening adjustment command to the PLC controller. The gate opening adjustment amount ΔG = min(0.5 × ΔS) c , 100mm), where ΔS c The vertical displacement of the sensor node that triggers the warning;

[0062] When the mobile terminal of the on-site maintenance personnel receives a level 2 or higher warning signal, it automatically activates the positioning and navigation function. The positioning and navigation function calls the coordinate data of the Beidou positioning module to generate the optimal path from the current location to the warning sensor node.

[0063] The special warning signal triggered in step S4 is synchronously transmitted to the maritime supervision platform through the communication base station, and the maritime supervision platform is notified to generate an electronic fence with a radius of 500 meters for the ship navigation restricted area based on the warning location coordinates.

[0064] All linkage control operations record the operation timestamp and execution parameters, and are transmitted back to the data processing center for closed-loop verification via the data acquisition module.

[0065] Preferably, in step S5, the early warning server sends the early warning signal to the terminal device through the communication base station using a dual-channel transmission mode of 4G network and Beidou short message, employing the following dynamic selection mechanism:

[0066] A signal quality monitoring unit is set up in the communication base station. The signal quality monitoring unit collects the RSSI value of the 4G network and the latency data of Beidou short messages in real time. When the RSSI is lower than -90dBm and the latency data exceeds 5 seconds, the forced handover logic is activated.

[0067] The forced switching logic performs the following operations: data packets containing red identification codes in the warning signals are preferentially transmitted using BeiDou short message service, while the remaining data packets are cached locally and retransmitted after the 4G signal is restored;

[0068] One hour before the daily peak tide, the communication base station automatically starts the dual-channel redundant transmission mode. In the redundant transmission mode, the level 3 and above warning signals are simultaneously sent twice via the 4G network and Beidou short message, with the interval between the two transmissions set to 30 seconds.

[0069] The data acquisition module's data caching unit records the actual number of successful transmissions for each data packet. When the same data packet fails to transmit in both channels, it automatically triggers the on-site audible and visual alarm and generates a communication fault log.

[0070] In step S4, the special warning signal generated by the settlement trend prediction module is accompanied by an emergency retransmission mark during transmission. When the communication base station receives the data packet with the emergency retransmission mark, if the first transmission fails, it will resend it at 2-minute intervals until it receives the confirmation receipt from the terminal device.

[0071] When the mobile terminal of the terminal device receives data in BeiDou short message mode, it automatically compares the data integrity of the 4G channel and the BeiDou channel. When the data loss rate is detected to be more than 20%, it actively initiates a retransmission request to the communication base station.

[0072] The present invention includes at least the following beneficial effects: (1) By deploying three-line sensors and fusing multi-source data, the vertical settlement, horizontal deformation and environmental interference of the dam can be monitored in a coordinated manner, which significantly improves the comprehensiveness and real-time performance of settlement analysis; (2) Based on stiffness parameter optimization and layered settlement identification model, the detection sensitivity of material performance degradation and deep soil anomalies is enhanced, and the accuracy of early warning is improved; (3) By eliminating tidal interference and temperature compensation mechanism, environmental factors and actual settlement are effectively separated, and the risk of misjudgment under complex working conditions is reduced; (4) By adopting dynamic threshold adjustment and geological parameter verification, a progressive assessment from local soil condition to overall structural safety is achieved, which improves the risk classification response capability; (5) By using emergency acquisition mode and dual-channel redundant transmission, the integrity of monitoring data and the reliable delivery of early warning instructions under extreme environments are ensured, and the robustness of the system is strengthened. Attached Figure Description

[0073] Figure 1 The principle flowchart of the coastal protection dike settlement monitoring method provided by the present invention. Detailed Implementation

[0074] The present invention will now be described in further detail with reference to the accompanying drawings, so that those skilled in the art can implement it based on the description.

[0075] like Figure 1 As shown, the coastal protection dike settlement monitoring method provided by the present invention includes the following steps:

[0076] S1: A longitudinal monitoring section is set on the leeward slope of the main body of the dike. The longitudinal monitoring section extends along the dike axis and includes a top survey line arranged on the top of the dike, a toe survey line arranged on the toe of the dike, and a middle survey line arranged on the middle platform of the dike. Fiber grating sensor arrays are set at intervals on the top survey line, the toe survey line and the middle survey line. The fiber grating sensor arrays include multiple sensor nodes arranged at intervals along the dike axis. Each sensor node includes a displacement meter arranged vertically and an inclination sensor arranged horizontally.

[0077] The seaward side of the dike is a sensitive area for settlement and deformation due to long-term seawater erosion and soil seepage. Selecting this area to set up monitoring sections can prioritize the capture of potential failure initiation points. In the longitudinal monitoring section structure, the dike crest measuring line is located in the middle of the dike crest to monitor the compressive deformation of the dike crest caused by its own weight and vehicle loads; the intermediate platform measuring line is arranged on the structural platform at 1 / 2 of the dike height to reflect the consolidation settlement characteristics of the soil in the middle of the dike; and the dike toe measuring line is arranged close to the dike toe line to monitor the scour settlement of the base caused by tidal changes.

[0078] The fiber optic grating sensor array is arranged with sensor nodes every 20 meters along the dam axis (the conventional interval can be adjusted to 10-30 meters depending on the length of the dam). The sensor types include vertical displacement gauges, which are vibrating wire displacement gauges that measure vertical displacement by changing the frequency of the steel wire (range ±50mm, resolution 0.01mm); and tilt sensors, which measure horizontal tilt angle based on MEMS technology (range ±15°, accuracy 0.01°).

[0079] The top survey line focuses on overall settlement, the middle survey line monitors the stratified settlement trend, and the toe survey line captures the base deformation, forming a three-dimensional monitoring network; the sensor nodes spaced 20 meters apart can balance monitoring accuracy and engineering cost, and avoid missing local deformation.

[0080] S2: A data acquisition module is set up in the monitoring and control station. The data acquisition module is connected to three sets of fiber optic grating sensor arrays on the longitudinal monitoring section through armored optical cables. The data acquisition module includes a synchronization trigger unit and a data buffer unit. The synchronization trigger unit periodically sends synchronization acquisition commands to all sensor nodes. The data buffer unit receives and stores the vertical displacement data measured by the displacement gauge and the horizontal tilt data measured by the tilt sensor.

[0081] Double-layered stainless steel armored optical cables connect the sensors and monitoring stations, with a tensile strength ≥2000N to prevent seawater corrosion and mechanical damage. The synchronization trigger unit uses periodic commands with a 1-hour synchronization acquisition cycle (standard selection, adjustable to 2 hours based on tidal cycles). A synchronization mechanism is also employed, sending pulse signals via the optical cable to ensure that the timestamp error of all node data is <10ms. The data buffer unit uses a ring buffer design, storing ≥72 hours of raw data, including vertical displacement data (16-bit precision), tilt angle data (floating-point), and timestamps (UTC format). The trigger unit sends a synchronization pulse every hour on the hour, and the sensor nodes complete displacement and tilt angle measurements within 10ms. Data is transmitted via optical cable to the buffer unit, where data packets are generated according to "survey line number + node location" to avoid data confusion.

[0082] S3: Configure a multi-source data fusion module in the data processing center. The multi-source data fusion module receives vertical displacement data and horizontal tilt angle data transmitted by the data acquisition module, establishes a joint analysis model of vertical displacement and horizontal tilt angle, and uses the Kalman filter algorithm to eliminate periodic displacement interference caused by tidal changes and generate a settlement distribution curve in the dam axis direction.

[0083] The joint analysis model architecture includes: Input layer: vertical displacement data (time series) and horizontal tilt data (spatial distribution); Processing layer: Kalman filtering algorithm to eliminate periodic tidal interference, and spatial interpolation algorithm to generate continuous settlement curves (such as Kriging interpolation); Output layer: output settlement distribution curves (horizontal axis is the distance from the dam axis, and vertical axis is the settlement amount).

[0084] After receiving the transmitted raw data, the data fusion module first aligns the timestamps; then corrects the directional deviation of the displacement measurement using tilt angle data (e.g., triggering displacement compensation when the tilt angle is >1°); and finally generates the net settlement distribution curve after Kalman filtering to remove tidal periodic interference.

[0085] S4: Deploy a settlement trend prediction module on the early warning server. The settlement trend prediction module connects to the multi-source data fusion module and obtains the settlement distribution curve. When the vertical displacement data of any sensor node changes for N consecutive cycles, the change amount exceeds the change threshold D, a level 3 early warning signal is triggered. When the vertical displacement change gradient of three adjacent sensor nodes exceeds the gradient threshold B, a level 2 early warning signal is triggered. When the radius of curvature of the settlement distribution curve is less than the design allowable value C, a level 1 early warning signal is triggered.

[0086] A three-level early warning system is adopted. The third-level early warning is for continuous settlement at a single point (e.g., N is 3 and D is 5mm, indicating that the cumulative settlement over 3 cycles is >15mm), indicating local soil softening. The second-level early warning is for settlement gradient exceeding the standard at adjacent nodes (e.g., B is 2mm / m), indicating the formation of a potential sliding surface. The first-level early warning condition is: radius of curvature < design value (e.g., C is 500m, this value refers to the industry requirements for dam structure design, that is, when the radius of curvature R is <500m or close to this condition, it is triggered), indicating the risk of overall instability.

[0087] The prediction module architecture performs real-time analysis, including updating the settlement curve and early warning status every hour, comparing historical data, and using construction period acceptance data as a benchmark. A Level 3 early warning is triggered when the settlement at a node reaches 5.2mm, 5.5mm, and 5.8mm for three consecutive hours; a Level 2 early warning is triggered if the settlement gradient between nodes within 50 meters upstream and downstream reaches 2.3mm / m; and a Level 1 early warning is triggered when the radius of curvature of the settlement curve for that segment drops to 480m (the design allowable value is 500m).

[0088] S5: The early warning server sends early warning signals to terminal devices through communication base stations. The communication base stations adopt a dual-channel transmission mode of 4G network and Beidou short message. The terminal devices include computer terminals of the dam management center and mobile terminals of on-site maintenance personnel.

[0089] The communication base station is configured with a 4G network for regular data transmission (bandwidth ≥ 10Mbps) and can use BeiDou short message service as an emergency channel (single transmission capacity 78 bytes, supporting location feedback). The terminal equipment includes computer terminals with professional monitoring software that displays settlement curves and warning levels; mobile terminals receive warning information and navigation paths via an app. After generating a warning signal, the warning server prioritizes sending it to the management center via the 4G network; if the 4G signal is interrupted (e.g., during typhoon weather), it automatically switches to BeiDou short message service to transmit coordinates and warning codes; after receiving the warning, on-site personnel's mobile terminals automatically navigate to the target node (error < 5m).

[0090] This method achieves the following through the above steps: Precise monitoring capability: Through the arrangement of three measuring lines and the fusion of multiple sensors, it enables the synchronous capture of vertical settlement and horizontal deformation, reducing the missed detection rate to below 5%; Anti-interference optimization: Kalman filtering effectively separates tidal period interference, improving the signal-to-noise ratio of settlement data to a very high level; Graded response efficiency: The three-level early warning mechanism shortens the emergency response time to a very short time, and the accuracy of major hazard identification is very high; Communication robustness: Dual-channel transmission ensures data reachability of over 99.5% in extreme environments.

[0091] In another technical solution, step S3, establishing the joint analysis model of vertical displacement and horizontal tilt angle includes the following steps:

[0092] The vertical displacement data is divided into discrete segments with 10-minute intervals according to the time series. The standard deviation of the displacement change rate and the slope of the linear fitting are calculated for each discrete segment to generate a sequence of vertical displacement characteristic parameters.

[0093] The horizontal tilt angle data is subjected to fast Fourier transform, and the energy ratio of the 0.1-0.5Hz frequency band is extracted as the dynamic feature of the tilt angle. When the energy ratio exceeds 60%, it is determined to be abnormal vibration interference.

[0094] Establish a correlation coefficient matrix between the vertical displacement characteristic parameter sequence and the tilt angle dynamic characteristic quantity. When the absolute value of the correlation coefficient is greater than 0.7, mark the vertical displacement data of the corresponding time period as the segment affected by tilt angle coupling.

[0095] In the Kalman filter algorithm, the measured value of the elastic modulus of the dam's main structure is introduced. The process noise covariance matrix of the filter is adjusted based on the product of the measured elastic modulus and the dynamic characteristic of the tilt angle. The adjustment formula is as follows: Where Q is the adjusted process noise covariance matrix, Q0 is the original process noise covariance matrix, E is the measured value of elastic modulus, and A is the dynamic characteristic of tilt angle;

[0096] The vertical displacement data of the section marked as being affected by tilt coupling is input into the adjusted Kalman filter to generate a settlement distribution curve that includes the influence of structural stiffness, while removing data points corresponding to abnormal vibration interference.

[0097] Vertical displacement data is divided into discrete segments at fixed intervals (e.g., 10-minute segments) according to the time series. By calculating the standard deviation of the displacement change rate and the slope of the linear fitting within each time segment, characteristic parameters reflecting settlement dynamics are extracted. The standard deviation is used to quantify the intensity of settlement fluctuations, and the linear slope characterizes the direction of settlement trends. After the horizontal tilt angle data undergoes a fast Fourier transform, the energy proportion in the 0.1-0.5Hz frequency band (the conventional vibration monitoring frequency band) is analyzed. When the energy in this frequency band exceeds 60% of the total energy (an adjustable threshold, which can be set to 50%-70% depending on geological conditions), it is determined to be abnormal vibration interference (such as non-settlement deformation caused by mechanical construction or wave impact). By comparing the temporal correlation between vertical displacement characteristics and tilt angle dynamic characteristics (when the absolute value of the correlation coefficient exceeds 0.7), the time periods affected by tilt angle coupling are marked to avoid misjudging structural torsion as settlement.

[0098] The measured elastic modulus of the dam (obtained through borehole sampling during construction) is incorporated into the Kalman filter algorithm, and the filter parameters are dynamically adjusted in conjunction with the dynamic characteristic of the inclination angle. The elastic modulus reflects the material's resistance to deformation. When the elastic modulus is high (e.g., greater than 50 MPa), the adjustment amplitude of the noise covariance matrix during the filtering process increases, enhancing the sensitivity to deformation of rigid structures. The adjusted filter can distinguish between displacement differences caused by soil creep (low frequency) and changes in structural stiffness (mid-to-high frequency), thereby generating a more accurate settlement distribution curve. During implementation, the elastic modulus data is updated regularly via the Internet of Things (IoT) to ensure that the model parameters are consistent with actual working conditions.

[0099] Data from sections marked as inclination coupling zones are input into an optimized filter, simultaneously removing abnormal vibration data points with excessive energy proportions (e.g., removing periods with energy proportions > 60%). When outputting settlement curves, stiffness indicators are added to areas affected by structural stiffness (e.g., concrete slope protection sections) to highlight the risk of material fatigue. For example, in seawall monitoring, this method reduced the proportion of inclination interference data caused by wave impact from 25% to 8%, and decreased settlement calculation errors to ±1.2 mm.

[0100] By accurately identifying non-settlement interference through frequency band energy analysis, the ability to identify non-settlement interference signals is significantly improved, enhancing data effectiveness. By integrating structural stiffness parameters, the sensitivity to monitoring changes in material properties is enhanced. Optimized settlement curve output provides clearer guidance for targeted maintenance.

[0101] In step S3, after generating the settlement distribution curve that includes the influence of structural stiffness, the following steps are also performed:

[0102] Based on the time periods corresponding to correlation coefficients with absolute values ​​greater than 0.7 in the correlation coefficient matrix, the vertical displacement characteristic parameter sequences of the top, middle, and toe survey lines of the embankment are extracted for the same time period, and the displacement compatibility index η among the three survey lines is calculated. , where σ top σ mid σ foot These are the standard deviations of the vertical displacement rates of the survey lines at the top, middle, and toes of the embankment, respectively.

[0103] The displacement compatibility index η is nonlinearly fitted with the measured value of the elastic modulus to establish an η-E relationship surface. When the curvature of the η-E relationship surface exceeds a preset threshold, it is determined that the main body of the dam has a tendency of layered settlement.

[0104] A stiffness anomaly detection unit is added to the settlement trend prediction module. The stiffness anomaly detection unit receives η-E relationship surface data. When a layered settlement trend is detected, a correction amount ΔR=γ×(η-η0) is automatically added to the curvature radius value of the corresponding embankment segment in the settlement distribution curve, where η0 is the benchmark coordination index during embankment construction and acceptance, and γ is the reciprocal of the width of the main cross-section of the embankment.

[0105] In step S4, the following judgment rule is added to the conditions for triggering the first-level warning signal: when the corrected radius of curvature value is less than the design allowable value and the curvature of the η-E relationship surface exceeds the preset threshold, the warning signal will be upgraded to the special warning signal.

[0106] When generating the settlement distribution curve, the multi-source data fusion module simultaneously outputs a three-dimensional deformation cloud map containing stratified settlement trend indicators. The distribution range of the η value is distinguished by color gradient in the three-dimensional deformation cloud map.

[0107] Based on the correlation coefficient matrix, highly correlated time periods (correlation coefficient > 0.7) are selected, and the vertical displacement characteristic parameters of the three survey lines (top, middle, and toe) are extracted during the same time period. The displacement compatibility index η is defined by calculating the ratio of the standard deviations of the displacement change rates of the three survey lines (e.g., the difference between the top and middle survey lines, and the difference between the middle and toe survey lines). The closer the η value is to 1, the better the deformation compatibility of the three survey lines; when the η value is below 0.5 (an adjustable threshold), it indicates a risk of stratified settlement. For example, in the monitoring of soft soil foundation embankments, if the η value drops sharply from 0.8 to 0.3, it corresponds to the accurate depth of the soil stripping layer detected by subsequent ground-penetrating radar.

[0108] The η value is nonlinearly fitted to the elastic modulus E (e.g., quadratic surface fitting) to establish an η-E relationship surface. When the surface curvature exceeds a preset threshold (e.g., curvature > 0.05 / m), a stratified settlement trend is identified. In implementation, the curvature threshold is trained using historical data to ensure adaptability to different geological conditions of different embankment sections. The stiffness anomaly detection unit adds a correction to the curvature radius of the settlement curve based on real-time η-E data (e.g., correcting the original curvature radius of 500m to 480m), making the early warning judgment closer to the actual structural state.

[0109] The generated 3D deformation cloud map visually displays the stratified settlement trend distribution through color gradients (e.g., red represents η < 0.4, green represents η > 0.8). When the corrected radius of curvature is simultaneously less than the design value (e.g., 500m) and the η-E curvature exceeds the standard, the Level 1 warning is upgraded to a Level 3 warning. In one engineering case, this method provided a 12-hour advance warning of a potential landslide, enabling the emergency repair team to reinforce the embankment toe in time.

[0110] Through multi-line collaborative analysis, the accuracy of identifying stratified settlement trends is significantly improved, and the ability to predict potential risks is enhanced; the η-E surface model realizes coupled early warning of stiffness and deformation, effectively reducing misjudgments caused by changes in structural stiffness and improving the reliability of early warning; and the deformation characteristics are intuitively displayed through three-dimensional visualization technology, which greatly improves the efficiency of decision support.

[0111] In step S3, after generating the three-dimensional deformation cloud map containing stratified settlement trend indicators, the following steps are also performed:

[0112] Among the top survey line, middle survey line and toe survey line of the dike, select the section where the sensor node with the lowest η value is located. Use a mobile detection vehicle to carry a ground-penetrating radar to scan the profile along the section. The ground-penetrating radar uses an 800MHz antenna to collect dielectric constant data at 0.2-meter intervals. At the same time, use a static cone penetration test device to measure the cone tip resistance value at a distance of 1 meter from the sensor node.

[0113] The dielectric constant data and cone tip resistance value are input into the multi-source data fusion module to establish a joint verification model of dielectric constant-cone tip resistance. The model outputs a correction coefficient μ for the compaction of the main soil mass of the dam. , where ε r q is a relative value of the dielectric constant. c This represents the cone tip resistance value.

[0114] The calculation formula of the displacement compatibility index η is adjusted according to the μ value, and is changed to η'=η×(1+0.2μ). The adjusted displacement compatibility index η' is then substituted back into the η-E relationship surface to determine the stratified settlement trend.

[0115] In step S2, the synchronization triggering unit of the data acquisition module dynamically adjusts the acquisition cycle according to the adjusted η' value: when η' < 0.6, the acquisition cycle is shortened to 30 minutes; when 0.6 ≤ η' ≤ 0.8, the cycle is maintained at 1 hour; when η' > 0.8, the acquisition cycle is extended to 2 hours.

[0116] In step S5, soil compaction correction information is added when the warning signal is sent. When μ < 1.2, a red identification code is added to the warning signal; when 1.2 ≤ μ ≤ 1.5, a yellow identification code is added; and when μ > 1.5, a green identification code is added.

[0117] The BeiDou short message transmission mode of the communication base station adjusts the transmission frequency according to the μ value: the red identification code corresponds to 3 times per hour, the yellow identification code corresponds to 1 time per hour, and the green identification code corresponds to 1 time every 3 hours.

[0118] A mobile detection vehicle equipped with a ground-penetrating radar (e.g., an 800MHz antenna) is deployed at the section with the lowest η value (typically the weakest area) to acquire the dielectric constant distribution through high-density scanning (0.2-meter intervals), reflecting changes in soil moisture content. Simultaneously, a static cone penetration tester (2cm / s) is used to measure the cone tip resistance, obtaining soil shear strength data. A joint verification model is then used to correlate the dielectric constant with the cone tip resistance (e.g., ε). r ×q c When the value is greater than 10, it is determined to be a dense area. The soil density correction coefficient μ is output to quantify the local soil stability.

[0119] The data acquisition cycle is dynamically adjusted based on the μ value: shortened to 30 minutes when μ < 1.2 (loose soil), and extended to 2 hours when μ > 1.8 (dense soil). Warning signals are accompanied by red, yellow, and green identification codes (e.g., red corresponds to μ < 1.2, requiring immediate action). Communication base stations adjust the BeiDou message transmission frequency according to the identification code (red 3 times per hour, green once every 3 hours). In reclamation dikes, this method increases the data acquisition frequency in high-risk areas by 3 times and reduces communication resource consumption by 40%.

[0120] After receiving the warning, the mobile terminal automatically navigates to the target node (positioning error < 3 meters) and displays the real-time μ value to assist in on-site judgment. For example, a yellow warning (μ=1.4) indicates that on-site verification reveals localized gravel accumulation, requiring no emergency response and avoiding overreaction. The density data is simultaneously fed back to the data processing center to update the regional weight parameters of the η-E relationship model.

[0121] On-site verification of geological parameters significantly improves the accuracy of on-site soil condition verification and enhances the ability to assess local risks; dynamically optimizes the data acquisition frequency to effectively balance monitoring accuracy and resource consumption; and achieves hierarchical response through an identification code system to improve the rationality of emergency resource allocation.

[0122] In another technical solution, step S3, the vertical displacement-horizontal tilt angle joint analysis model further includes the following processing steps:

[0123] The data processing center receives real-time tide data transmitted from the tide monitoring station. The tide monitoring station is located on the seaward side of the main body of the dike and 10 meters away from the toe of the dike. The tide monitoring station includes a pressure level gauge and a Beidou positioning module. The pressure level gauge measures the tide elevation in 5-minute intervals, and the Beidou positioning module records the latitude and longitude coordinates of the tide monitoring station.

[0124] The multi-source data fusion module timestamps and matches the tide level elevation with the vertical displacement data. It then uses a Kalman filter algorithm to extract the periodic displacement component caused by the tide level change. The extraction frequency of the periodic displacement component is 0.5-2 times / hour, and this component is removed from the vertical displacement data.

[0125] Based on the correlation analysis results between horizontal dip angle data and tidal level elevation, the weighting coefficients of the dip angle sensors in the joint analysis model are dynamically adjusted. The adjustment formula for the weighting coefficient α is as follows: , where ΔH is the difference in elevation between two adjacent tides, and k is the proportional factor corresponding to the permeability coefficient of the main soil of the dam.

[0126] In step S4, when the settlement trend prediction module triggers the secondary early warning signal, it simultaneously performs the following operations:

[0127] Extract the tide phase data of the tide monitoring station in the 24 hours before the current time, calculate the difference between the actual tide phase and the historical tide phase for the same period, and upgrade the level II warning signal to the level I warning signal when the difference exceeds 30 degrees and the vertical displacement change gradient exceeds 2 mm / m.

[0128] After generating the settlement distribution curve, based on the time delay correlation between the tide level data and the vertical displacement data, the displacement compensation of the sensor nodes within 200 meters of the tide level monitoring station in the settlement distribution curve is corrected. The displacement compensation is as follows: H inst H represents the current tide level elevation. mean β is the average tide level for the month, and β is the tangent of the inclination angle of the seaward slope of the main body of the dike.

[0129] A tide level monitoring station is deployed 10 meters from the toe of the dike on the seaward side (the location selection takes into account both the area directly affected by tides and equipment safety). The station is equipped with a pressure-type water level gauge and a BeiDou positioning module. The pressure-type water level gauge calculates the tide level elevation by measuring water pressure (range 0-10 meters, accuracy ±1 cm), collecting data every 5 minutes (adjustable to 2-10 minutes to adapt to the rate of tidal changes). The BeiDou module records the latitude and longitude coordinates of the monitoring station (positioning accuracy ±0.5 meters), ensuring accurate correlation between tide level data and dike location. Data is transmitted wirelessly in real-time to the data processing center, aligned with the timestamps of vertical displacement data (error controlled within ±10 seconds). For example, in estuary dikes, the synchronization deviation between tide level data and displacement data can be reduced from 30 seconds to 8 seconds, significantly improving the reliability of correlation analysis.

[0130] The multi-source data fusion module separates the periodic influence of tidal levels using a Kalman filter algorithm. The algorithm identifies the displacement component caused by tidal changes (frequency range 0.5-2 times / hour, corresponding to semi-diurnal to diurnal tide characteristics) and removes it from the original displacement data. Simultaneously, it dynamically adjusts the weighting coefficient of the tilt sensor based on the tidal elevation difference (e.g., ΔH = 0.8 meters between two adjacent tidal changes). The weighting coefficient is related to the permeability coefficient of the dam soil (e.g., k = 0.3 for sandy soil, k = 0.1 for clay); the higher the soil permeability, the higher the weight of the tidal change's influence on the tilt angle. In practice, when ΔH exceeds 1 meter (e.g., during storm surges), the weighting coefficient automatically increases to 0.9, enhancing the sensitivity of the tilt angle data to structural tilt.

[0131] When a Level II warning is triggered (settlement gradient exceeding the standard at three adjacent nodes), the tidal phase difference over 24 hours is analyzed simultaneously. If the actual tidal phase deviates from the historical average by more than 30 degrees (adjustable threshold, such as 20-40 degrees), and the settlement gradient continues to exceed the standard, the warning is upgraded to Level I. For example, during a typhoon, the tidal phase deviation can reach 45 degrees. Combined with a settlement gradient of 2.5 mm / m, the system triggers a Level I warning in advance to prevent the expansion of cracks in the dike crest. In addition, for sensor nodes within 200 meters of the tidal monitoring station (areas significantly affected by tidal erosion), the displacement compensation is calculated based on the difference between the current tidal level and the monthly average. In the compensation calculation, the slope inclination tangent β is obtained from the design drawings (e.g., β=0.33 for a 1:3 slope ratio), and after correction, the interference of short-term tidal fluctuations on the long-term settlement trend is eliminated.

[0132] Tidal data fusion effectively eliminates tidal periodic interference and improves the accuracy of settlement data analysis; the dynamic weight adjustment mechanism dynamically adjusts sensor weights to enhance monitoring adaptability under complex working conditions; the phase difference correlation early warning upgrade significantly improves the collaborative early warning capability for complex disasters through the phase correlation mechanism, gaining critical time windows for emergency repair decisions.

[0133] In another technical solution, in step S1, the sensor nodes of the fiber Bragg grating sensor array further include the following structural features:

[0134] A temperature compensation unit is set between the displacement gauge and the tilt sensor at each sensor node. The temperature compensation unit includes a platinum resistance temperature sensor that is in close contact with the surface of the dam body. The platinum resistance temperature sensor measures the contact temperature at the sensor node with a resolution of 0.1℃.

[0135] The output of the temperature compensation unit is connected to the data acquisition module, and the data buffer unit synchronously stores the temperature measurement data. The temperature measurement data has the same timestamp as the vertical displacement data and the horizontal tilt angle data.

[0136] In step S3, when establishing the vertical displacement-horizontal tilt angle joint analysis model, the multi-source data fusion module inputs temperature measurement data into the displacement compensation model. The displacement compensation model corrects the vertical displacement data in real time based on the temperature-displacement relationship curve. The formula for calculating the vertical displacement correction is as follows: , where α T T is the temperature sensitivity coefficient of the displacement gauge. inst The current temperature measurement value, T ref This is the reference temperature value during installation and commissioning;

[0137] Before processing the vertical displacement data, the Kalman filter algorithm first applies the corrected ΔD... c Accumulated amounts are deducted, and the deduction period is consistent with the 1-hour period of the synchronous acquisition command.

[0138] A temperature compensation unit is integrated between the displacement gauge and tilt sensor at each sensor node. This unit uses a platinum resistance temperature sensor (PT100 type, temperature range -20℃~80℃) installed close to the dam surface to measure the contact temperature in real time with a resolution of 0.1℃. Temperature data is transmitted to the data acquisition module via an independent channel, sharing the same timestamp (error <10ms) with the displacement and tilt data, ensuring synchronous analysis of multiple parameters. For example, in monitoring seawalls in cold regions, the temperature sensor detected a diurnal temperature difference of up to 25℃, and the synchronous displacement data fluctuated by 1.8mm, verifying the necessity of temperature compensation. The sensor node housing adopts a waterproof and sealed design (IP68 protection rating) to prevent seawater penetration that could cause temperature measurement distortion.

[0139] The multi-source data fusion module calls temperature measurement data and performs real-time correction on the original vertical displacement based on a preset temperature-displacement relationship curve (obtained through laboratory calibration). The correction amount calculation depends on the temperature sensitivity coefficient of the displacement gauge (e.g., α of a certain model of displacement gauge). T =0.02mm / ℃), when the deviation between the site temperature and the installation reference temperature (usually the annual average temperature) exceeds ±5℃, the compensation calculation is automatically triggered. In practice, the data processing center deducts the cumulative correction amount (e.g., a cumulative correction of 1.0mm over 5 consecutive hours) every hour to avoid data jumps caused by point-by-point correction. This method reduces the displacement measurement error under low-temperature conditions in winter from ±1.5mm to ±0.3mm.

[0140] Before formal processing, the Kalman filter algorithm preprocesses the temperature-corrected vertical displacement data. This preprocessing includes: eliminating linear drift caused by temperature (e.g., deducting a cumulative amount of 0.02 mm / ℃ per hour); and marking data during periods of abrupt temperature changes (e.g., ΔT > 8℃ within 1 hour) for subsequent analysis. In levees in tropical regions, it can successfully identify false settlement alarms caused by sudden temperature drops due to heavy rain (ΔT = 12℃ in 2 hours), avoiding false triggering of Level III warnings.

[0141] Temperature compensation significantly improves the accuracy of displacement measurement under varying temperature conditions and reduces the impact of environmental interference; the preprocessing mechanism effectively reduces false alarms caused by sudden temperature changes and improves system stability; multi-parameter timestamp alignment and multi-parameter synchronous analysis enhance the reliability of data correlation verification.

[0142] In step S4, the early warning judgment of the settlement trend prediction module also includes the following analysis process:

[0143] Extract the current temperature measurement value T recorded by the temperature compensation unit. inst Compared with the reference temperature value T ref The temperature anomaly correction mode is activated when the absolute value of ΔT exceeds 5°C for three consecutive cycles.

[0144] In the temperature anomaly correction mode, a temperature compensation verification rule is added to the conditions for triggering the level 3 warning signal: when the vertical displacement data changes by more than 5 mm for three consecutive cycles, ΔD is calculated synchronously. c Compared with the measured displacement change ΔD measured ratio If the ratio K > 0.3, the warning signal will be downgraded to an observation-level alarm.

[0145] When determining whether the vertical displacement gradient of three adjacent sensor nodes exceeds 2 mm / m, the standard deviation σ of ΔT at these three nodes is calculated simultaneously. T When σ TWhen the temperature is >2℃, the gradient threshold is adjusted from 2 mm / m to 2.5 mm / m;

[0146] When calculating the radius of curvature of the settlement distribution curve, a cumulative temperature influence factor is introduced. , where ΔT i This represents the hourly ΔT value within the previous 24 hours of the current calculation period, where n is the number of samples. The design allowable value is multiplied by a correction factor [1 + 0.02 × (C)]. T -10)];

[0147] In step S5, the warning signal is accompanied by a temperature-related identifier code. When the absolute value of ΔT exceeds 10℃, a temperature anomaly marker is embedded in the warning signal. This marker triggers the mobile terminal of the terminal device to start the temperature calibration mode. The temperature calibration mode forcibly calls the ground temperature monitoring data of the most recent 3 hours and cross-verifies it with the data of the platinum resistance temperature sensor.

[0148] When the absolute value of the temperature deviation ΔT exceeds 5℃ (adjustable threshold, set to 3-8℃ depending on the regional climate) for three consecutive data collection cycles (which can be set from 1 to 5 cycles), the system activates the temperature anomaly correction mode. In this mode, triggering a Level 3 warning requires additional verification of the temperature correction amount ΔD. c The ratio K to the measured displacement change. For example, when ΔD c If the percentage exceeds 30% (i.e., K>0.3), it indicates that the current displacement change is mainly caused by temperature expansion, and the warning will be downgraded to an observation-level alarm (only recorded, without triggering an emergency response). A certain seawall has used this rule to reduce the number of false alarms per day from more than ten to two or fewer during the high-temperature summer season.

[0149] When calculating the settlement gradient of three adjacent nodes, the standard deviation σ of the temperature deviation in this region is analyzed simultaneously. T When σ T When the temperature exceeds 2℃ (which can be set to 1.5-3℃), it indicates that uneven temperature field distribution may amplify local settlement differences. In this case, the original gradient threshold of 2mm / m is relaxed to 2.5mm / m (the adjustment range can increase with the increase of σ_T). For example, if σ_T is at three nodes during monitoring... T =2.8℃, the system automatically adopts a threshold of 2.6mm / m to avoid misjudgment caused by uneven sunlight exposure.

[0150] Temperature accumulation effect compensation and terminal response mechanism introduce temperature accumulation influence factor C T (24-hour ΔT mean) Corrected design allowable value. When C TWhen the temperature exceeds 10℃ (e.g., during prolonged periods of high temperatures), the design allowable amplification factor increases by 0.02 / ℃ to accommodate the thermal expansion effect of materials. After the warning signal is embedded with a temperature anomaly marker, the mobile terminal automatically initiates calibration mode: it calls data from nearby ground temperature monitoring stations (if available) for cross-validation. It displays the percentage of temperature impact (e.g., temperature contribution to the current displacement is 45%) for on-site assessment. During cold wave events, this mechanism helps engineers quickly distinguish between true settlement (temperature contribution <20%) and frost heave deformation (temperature contribution >60%).

[0151] Temperature-related early warning rules significantly optimize the early warning logic in abnormal temperature scenarios, reducing unnecessary emergency responses; dynamic gradient threshold mechanism improves the monitoring adaptability of complex temperature fields; and temperature contribution analysis enhances the scientific nature of on-site response decisions.

[0152] In another technical solution, in step S2, the synchronization triggering unit further includes a dynamic acquisition mode switching function, which operates as follows:

[0153] The data acquisition module continuously monitors the rate of change of the vertical displacement data. When the vertical displacement change of any sensor node exceeds 3 mm for two consecutive cycles, it automatically switches to emergency acquisition mode.

[0154] In emergency data acquisition mode, the synchronous triggering unit shortens the acquisition cycle to 10 minutes and prioritizes the transmission of data from all nodes within a 50-meter range upstream and downstream of the sensor node.

[0155] In emergency acquisition mode, the data caching unit activates a block storage mechanism, marking priority data as high-priority data packets. High-priority data packets are transmitted in real time through the 4G network channel, while the remaining data is stored in the local cache.

[0156] In step S4, when the settlement trend prediction module receives a high-priority data packet, it starts a fast prediction algorithm. The fast prediction algorithm uses the exponential smoothing method to extrapolate the data of the most recent 6 periods. When the vertical displacement acceleration of the extrapolation result exceeds 0.05 mm / min², a secondary warning signal is forcibly triggered.

[0157] The BeiDou short message transmission mode of the communication base station is switched to continuous transmission mode in emergency collection mode, and the continuous transmission interval is set to 5 minutes.

[0158] When the vertical displacement change of any sensor node exceeds 3 mm (adjustable threshold, set to 2-5 mm depending on the geological risk level) for two consecutive acquisition cycles (configurable to 1-3 cycles), the synchronous trigger unit automatically switches to emergency acquisition mode. In this mode, the acquisition cycle is shortened from the usual 1 hour to 10 minutes (adjustable range 5-15 minutes), and data from all sensors within 50 meters upstream and downstream of the abnormal node is prioritized for transmission (covering the potentially affected area). For example, in the early stages of a piping failure, the system increases the acquisition frequency by 6 times within 2 hours to promptly detect the accelerated subsidence trend of 5 surrounding nodes. The data caching unit activates a block storage mechanism, transmitting high-priority data packets (marked in red) in real time via the 4G network, while regular data is temporarily stored locally (for up to 24 hours) to ensure that critical information is not lost.

[0159] Upon receiving high-priority data packets, the settlement trend prediction module immediately initiates exponential smoothing for extrapolation calculations. The algorithm calculates vertical displacement acceleration based on the most recent six cycles (i.e., six 10-minute data points within one hour). For example, a value exceeding the threshold of 0.05 mm / min² (e.g., 0.06 mm / min²) is used. If the predicted value continues to exceed the threshold, a level-two warning is forcibly triggered (skipping the regular analysis process). Simultaneously, BeiDou short message transmission switches to continuous transmission mode (5-minute interval, adjustable to 3-10 minutes), ensuring at least 12 transmissions of critical data per hour even under extreme network interruptions. During typhoons, this mechanism can reduce the warning response time from the usual 45 minutes to 18 minutes.

[0160] During high-priority data packet transmission, the communication base station automatically allocates 80% of its bandwidth resources (adjustable ratio) to ensure priority passage. All emergency operation records (such as mode switching time and extrapolated parameters) are appended to the data packet metadata for post-event analysis. When communication is restored, the system automatically retransmits locally cached data and optimizes algorithm parameters (e.g., adjusting the exponential smoothing coefficient) by comparing predicted and actual values.

[0161] The emergency mode significantly improves the data capture density in the early stages of a crisis, shortens the emergency response window, and buys valuable time for decision-making; it accelerates the identification of major risks through a mandatory early warning mechanism; and it optimizes communication resource allocation through dynamic bandwidth allocation to ensure the timeliness of critical data transmission.

[0162] In another technical solution, step S5, the warning signal processing of the terminal device further includes the following linkage control logic:

[0163] When the computer terminal of the dam management center receives a Level 1 early warning signal, it automatically activates the emergency control interface. This interface connects to the PLC controller of the dam's drainage gate, sending a gate opening adjustment command to the PLC controller. The gate opening adjustment amount ΔG = min(0.5 × ΔS) c , 100mm), where ΔSc The vertical displacement of the sensor node that triggers the warning;

[0164] When the mobile terminal of the on-site maintenance personnel receives a level 2 or higher warning signal, it automatically activates the positioning and navigation function. The positioning and navigation function calls the coordinate data of the Beidou positioning module to generate the optimal path from the current location to the warning sensor node.

[0165] The special warning signal triggered in step S4 is synchronously transmitted to the maritime supervision platform through the communication base station, and the maritime supervision platform is notified to generate an electronic fence with a radius of 500 meters for the ship navigation restricted area based on the warning location coordinates.

[0166] All linkage control operations record the operation timestamp and execution parameters, and are transmitted back to the data processing center for closed-loop verification via the data acquisition module.

[0167] Upon receiving a Level 1 warning signal, the computer terminal at the dam management center immediately activates the emergency control interface. This interface connects to the PLC controller of the drainage gate via the Modbus protocol and sends an opening adjustment command. The adjustment amount ΔG is calculated based on the node displacement ΔS that triggered the warning (e.g., ΔG = 10mm when ΔS = 20mm), ΔG = min(0.5 × ΔS). c (100mm) indicates that the adjustment amount is 0.5×ΔS c The smaller amount within 100mm is adjusted, and a maximum adjustment limit of 100mm is set to prevent excessive leakage. Simultaneously, the mobile terminals of on-site maintenance personnel (equipped with a customized app) automatically activate the BeiDou navigation function, generating the optimal path based on the coordinates of the warning nodes (avoiding landslide areas, with a path planning error of <3 meters). In a practical exercise, this function reduced the arrival time of the repair team from an average of 25 minutes to 12 minutes.

[0168] The highest-level early warning signal is simultaneously transmitted to the maritime monitoring platform via communication base stations. After analyzing the warning coordinates, the platform automatically generates an electronic no-navigation zone with a radius of 500 meters (adjustable to 300-800 meters). The no-navigation order is broadcast to surrounding vessels via the AIS system, and the channel indicator lights are switched to red for warning. For example, after a landslide warning is triggered, the electronic fence can successfully intercept fishing boats that have mistakenly entered the danger zone, preventing secondary accidents.

[0169] All coordinated operations (such as gate opening, navigation path, and navigation restriction instructions) are recorded with precise timestamps and execution parameters (e.g., actual gate movement delay <2 seconds). The data acquisition module monitors the operational effects in real time (e.g., the displacement change rate of the corresponding embankment section after drainage) and sends the feedback data back to the processing center for effect verification. If the operation does not meet expectations (e.g., displacement does not decrease within 1 hour), the system automatically upgrades the warning level or switches the response plan.

[0170] Automatic gate adjustment significantly improves the level of automation in hazard control and accelerates the efficiency of emergency operations; electronic fences effectively reduce the risk of secondary disasters and enhance the collaborative response capabilities of multiple departments; closed-loop verification mechanisms improve the effectiveness and traceability of response measures.

[0171] In step S5, the early warning server sends an early warning signal to the terminal device through a communication base station, using a dual-channel transmission mode of 4G network and BeiDou short message, and employs the following dynamic selection mechanism:

[0172] A signal quality monitoring unit is set up in the communication base station. The signal quality monitoring unit collects the RSSI value of the 4G network and the latency data of Beidou short messages in real time. When the RSSI is lower than -90dBm and the latency data exceeds 5 seconds, the forced handover logic is activated.

[0173] The forced switching logic performs the following operations: data packets containing red identification codes in the warning signals are preferentially transmitted using BeiDou short message service, while the remaining data packets are cached locally and retransmitted after the 4G signal is restored;

[0174] One hour before the daily peak tide, the communication base station automatically starts the dual-channel redundant transmission mode. In the redundant transmission mode, the level 3 and above warning signals are simultaneously sent twice via the 4G network and Beidou short message, with the interval between the two transmissions set to 30 seconds.

[0175] The data acquisition module's data caching unit records the actual number of successful transmissions for each data packet. When the same data packet fails to transmit in both channels, it automatically triggers the on-site audible and visual alarm and generates a communication fault log.

[0176] In step S4, the special warning signal generated by the settlement trend prediction module is accompanied by an emergency retransmission mark during transmission. When the communication base station receives the data packet with the emergency retransmission mark, if the first transmission fails, it will resend it at 2-minute intervals until it receives the confirmation receipt from the terminal device.

[0177] When the mobile terminal of the terminal device receives data in BeiDou short message mode, it automatically compares the data integrity of the 4G channel and the BeiDou channel. When the data loss rate is detected to be more than 20%, it actively initiates a retransmission request to the communication base station.

[0178] The communication base station has a built-in signal quality monitoring unit that collects 4G signal strength (RSSI) and BeiDou message latency in real time. When the RSSI is below -90dBm (can be set to -85 to -100dBm to adapt to different devices) and the latency exceeds 5 seconds (adjustable to 3-8 seconds), a forced handover logic is triggered: red identification code data (such as high-risk warnings with μ<1.2) are prioritized for transmission via BeiDou (single packet size ≤78 bytes), and the remaining data is temporarily stored locally (for a maximum of 12 hours). One hour before the daily high tide peak (can be set in advance), the base station starts a dual-channel redundant transmission mode, repeatedly sending warnings of level three and above twice (with a 30-second interval) to ensure at least one successful reception.

[0179] The data buffer unit records the actual number of times each data packet is sent and its success status. When the same data packet fails on both channels (e.g., three consecutive unresponsive transmissions), an on-site audible and visual alarm (volume ≥ 90 dB, flashing frequency 2 Hz) is immediately triggered, and a fault log with location information is generated. For high-priority warning data packets, an emergency retransmission flag (highest priority) is added. If the first transmission fails, it is retransmitted at 2-minute intervals (adjustable to 1-5 minutes) until a terminal acknowledgment (ACK signal) is received. In the event of a fiber optic cable outage, this mechanism ensures that high-priority warning data packets are successfully delivered within 15 minutes, while the recovery time for a conventional 4G channel is up to 6 hours.

[0180] When a mobile terminal receives data in BeiDou mode, it automatically compares the data integrity between the 4G and BeiDou channels (e.g., verifying packet sequence number continuity). When the missing rate exceeds 20% (can be set to 15-25%), the terminal proactively initiates a retransmission request to the base station (carrying a list of missing packet IDs). The retransmission request is sent via BeiDou messages (occupying an independent channel), and the base station retransmits the data according to priority. Simultaneously, the terminal activates the offline caching function (storage capacity ≥ 48 hours of data) to ensure that critical information can be retrieved during network outages.

[0181] Intelligent channel switching significantly enhances communication robustness in extreme environments, ensuring the accessibility of critical information; intelligent retransmission mechanisms improve the reliability of high-level early warning transmission; and optimized terminal self-verification functions reduce the interference of missing data on decision-making.

[0182] It should be noted that although the steps are described in a specific order above, this does not mean that they must be performed in that order. In fact, some of these steps can be executed concurrently, or even in a different order, as long as the required functionality is achieved. The number of devices and processing scale described herein are for simplification of the invention; applications, modifications, and variations of this invention will be readily apparent to those skilled in the art.

[0183] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A method of monitoring settlement of a coastal protection bank, characterised in that, The method comprises the following steps: S1: a longitudinal monitoring section is arranged on the seaward side slope surface of the dam body, the longitudinal monitoring section extends along the dam axis, the longitudinal monitoring section comprises a crest survey line arranged on the dam crest, a toe survey line arranged on the dam toe, and an intermediate survey line arranged on the intermediate platform of the dam body, the crest survey line, the toe survey line and the intermediate survey line are respectively provided with an optical fiber grating sensor array, the optical fiber grating sensor array comprises a plurality of sensor nodes arranged at intervals along the direction of the dam axis, each sensor node comprises a displacement meter arranged in the vertical direction and an inclination sensor arranged in the horizontal direction; S2: a data acquisition module is arranged in the monitoring control station, the data acquisition module is connected with the three groups of optical fiber grating sensor arrays of the longitudinal monitoring section through armored optical cables, the data acquisition module comprises a synchronous triggering unit and a data caching unit, the synchronous triggering unit periodically sends a synchronous acquisition instruction to all sensor nodes, and the data caching unit receives and stores the vertical displacement data measured by the displacement meter and the horizontal inclination data measured by the inclination sensor; S3: a multi-source data fusion module is configured in the data processing center, the multi-source data fusion module receives the vertical displacement data and the horizontal inclination data transmitted by the data acquisition module, establishes a vertical displacement-horizontal inclination joint analysis model, the model eliminates the periodic displacement interference caused by the tidal level change through a Kalman filter algorithm, and generates a settlement distribution curve in the direction of the dam axis; S4: a settlement trend prediction module is deployed on the early warning server, the settlement trend prediction module is connected with the multi-source data fusion module and obtains the settlement distribution curve, when the vertical displacement data of any sensor node changes by more than a change threshold D for N consecutive periods, a three-level early warning signal is triggered, when the vertical displacement change gradient of adjacent three sensor nodes exceeds a gradient threshold B, a two-level early warning signal is triggered, and when the curvature radius value of the settlement distribution curve is less than a design allowable value C, a one-level early warning signal is triggered; S5: the early warning server sends an early warning signal to a terminal device through a communication base station, the communication base station adopts a 4G network and a Beidou short message dual-channel transmission mode, and the terminal device comprises a computer terminal of a dam management center and a mobile terminal of a field maintenance personnel.

2. The coastal protection barrier settlement monitoring method of claim 1, wherein, In the step S3, the establishment of the vertical displacement-horizontal inclination joint analysis model comprises the following steps: The vertical displacement data is divided into discrete segments with an interval of 10 minutes according to the time sequence, the standard deviation of the displacement change rate and the linear fitting slope of each discrete segment are calculated, and a vertical displacement characteristic parameter sequence is generated; The horizontal inclination data is subjected to a fast Fourier transform, and the energy proportion of the 0.1-0.5 Hz frequency band is extracted as an inclination dynamic characteristic quantity, and when the energy proportion exceeds 60%, it is determined that there is abnormal vibration interference; A correlation coefficient matrix of the vertical displacement characteristic parameter sequence and the inclination dynamic characteristic quantity is established, when the absolute value of the correlation coefficient is greater than 0.7, the vertical displacement data of the corresponding period is marked as a section affected by the inclination coupling effect; In the Kalman filtering algorithm, the measured value of the elastic modulus of the dam main structure is introduced, and the process noise covariance matrix of the filter is adjusted according to the product of the measured value of the elastic modulus and the inclination dynamic characteristic quantity, and the adjustment formula is: Wherein, Q is the adjusted process noise covariance matrix, Q0 is the original process noise covariance matrix, wherein E is the measured value of the elastic modulus, and A is the inclination dynamic characteristic quantity; The vertical displacement data marked as the section affected by the inclination coupling effect is input into an adjusted Kalman filter, a settlement distribution curve containing the influence of the structural stiffness is generated, and the data points corresponding to the abnormal vibration interference are removed.

3. The coastal protection barrier settlement monitoring method of claim 2, wherein, The step S3, after generating the settlement distribution curve containing the structural stiffness influence, further performs the following steps: According to the time period corresponding to the correlation coefficient with an absolute value greater than 0.7 in the correlation coefficient matrix, the vertical displacement characteristic parameter sequences of the crest line, the middle line and the toe line at the same time period are extracted, and a displacement coordination index η between the three lines is calculated, wherein σ top , σ mid , and σ foot are the vertical displacement change rate standard deviations of the crest line, the middle line and the toe line, respectively. The displacement compatibility index η is nonlinearly fitted with the measured value of the elastic modulus to establish an η-E relationship surface, and when the curvature of the η-E relationship surface exceeds a preset threshold, it is determined that the dam main body has a layered settlement trend; In the settlement trend prediction module, a stiffness anomaly detection unit is added, which receives η-E relationship surface data, and when a layered settlement trend is detected, an automatic correction amount ΔR=γ×(η-η0) is added to the curvature radius value of the corresponding embankment section in the settlement distribution curve, where η0 is the baseline compatibility index at the dam construction acceptance, and γ is the inverse of the width of the dam main body cross section; In the step S4, the condition for triggering a first-level warning signal is added with the following determination rule: when the corrected curvature radius value is less than the design allowable value and the curvature of the η-E relationship surface exceeds the preset threshold, the warning signal is upgraded to a special-level warning signal; The multi-source data fusion module synchronously outputs a three-dimensional deformation cloud chart containing a layered settlement trend identifier when generating the settlement distribution curve, and the three-dimensional deformation cloud chart uses color gradients to distinguish the distribution range of η values.

4. The coastal protection barrier settlement monitoring method of claim 3, wherein, In the step S3, after generating the three-dimensional deformation cloud chart containing the layered settlement trend identifier, the following steps are further performed: In the crest survey line, the middle survey line, and the embankment toe survey line, the section where the sensor node with the lowest η value is located is selected, and a geological radar is used to perform profile scanning along the section by moving a detection vehicle, the geological radar uses an 800MHz antenna to collect dielectric constant data at 0.2m intervals, and a static cone penetration device is used to measure the cone tip resistance value 1m away from the sensor node; The dielectric constant data and the cone tip resistance value are input into the multi-source data fusion module to establish a dielectric constant-cone tip resistance joint verification model, and a dam main body soil body density correction coefficient μ is output by the model, wherein ε r is a dielectric constant relative value, q c is a cone tip resistance value; According to the μ value, the calculation formula of the displacement compatibility index η is adjusted to η'=η×(1+0.2μ), and the adjusted displacement compatibility index η' is re-substituted into the η-E relationship surface for layered settlement trend determination; In the step S2, the synchronous triggering unit of the data acquisition module dynamically adjusts the acquisition period according to the adjusted η' value: when η'<0.6, the acquisition period is shortened to 30 minutes; when 0.6≤η'≤0.8, the 1-hour period is maintained; when η'>0.8, the acquisition period is extended to 2 hours; In the step S5, soil density correction information is added when sending a warning signal: when μ<1.2, a red identification code is added to the warning signal; when 1.2≤μ≤1.5, a yellow identification code is added; and when μ>1.5, a green identification code is added; The Beidou short message transmission mode of the communication base station adjusts the sending frequency according to the μ value: the red identification code corresponds to 3 times per hour, the yellow identification code corresponds to 1 time per hour, and the green identification code corresponds to 1 time every 3 hours.

5. The coastal protection barrier settlement monitoring method of claim 1, wherein, In the step S3, the vertical displacement-horizontal inclination joint analysis model further includes the following processing process: Access the tidal level data transmitted by the tidal level monitoring station in real time at the data processing center, the tidal level monitoring station is located on the seashore side of the dam body and 10 meters away from the dam toe measuring line, the tidal level monitoring station comprises a pressure type water level gauge and a Beidou positioning module, the pressure type water level gauge measures the tidal level elevation at a period of 5 minutes, and the Beidou positioning module records the longitude and latitude coordinates of the tidal level monitoring station; The multi-source data fusion module synchronously matches the tidal level elevation with the vertical displacement data by time stamp, extracts the periodic displacement component caused by the change of the tidal level by a Kalman filtering algorithm, the extraction frequency range of the periodic displacement component is 0.5-2 times / hour, and the component is removed from the vertical displacement data; According to the correlation analysis results of the horizontal inclination data and the tidal elevation, the weight coefficient of the inclination sensor in the joint analysis model is dynamically adjusted, and the weight coefficient α adjustment formula is: where ΔH is the difference between the adjacent two tidal elevations, and k is the proportional factor corresponding to the dam main soil permeability coefficient. In the step S4, the settlement trend prediction module synchronously performs the following operations when triggering the secondary warning signal: Extract the tidal level phase data of the tidal level monitoring station within 24 hours before the current time, calculate the difference between the actual tidal level phase and the historical tidal level phase, when the difference exceeds 30 degrees and the vertical displacement change gradient exceeds 2 mm / m, the secondary warning signal is upgraded to the primary warning signal; After generating the settlement distribution curve, based on the time delay correlation between the tide level data and the vertical displacement data, the displacement compensation of the sensor nodes within 200 meters of the tide level monitoring station in the settlement distribution curve is corrected. The displacement compensation is as follows: H inst H represents the current tide level elevation. mean β is the average tide level for the month, and β is the tangent of the inclination angle of the seaward slope of the main body of the dike.

6. The coastal protection barrier settlement monitoring method of claim 1, wherein, In the step S1, the sensor node of the fiber grating sensor array further comprises the following structural features: A temperature compensation unit is arranged between the displacement meter and the inclination sensor of each sensor node, the temperature compensation unit comprises a platinum resistance temperature sensor close to the surface of the dam body, and the platinum resistance temperature sensor measures the contact temperature at the sensor node at a resolution of 0.1℃; The output end of the temperature compensation unit is connected to the data acquisition module, and the temperature measurement data is synchronously stored in the data buffer unit, and the temperature measurement data, the vertical displacement data and the horizontal inclination data have the same time stamp; In the step S3, the multi-source data fusion module inputs the temperature measurement data into the displacement compensation model when establishing the vertical displacement-horizontal inclination joint analysis model, and the displacement compensation model corrects the vertical displacement data in real time according to the temperature-displacement relationship curve, and the vertical displacement correction amount calculation formula is: Wherein, α T is the temperature sensitive coefficient of the displacement meter, T inst is the current temperature measurement value, T ref is the reference temperature value during installation and debugging; The Kalman filter algorithm first carries out accumulation deduction on the corrected ΔD c before processing the vertical displacement data. The deduction period is consistent with the 1-hour period of the synchronous acquisition instruction.

7. The coastal protection barrier settlement monitoring method of claim 6, wherein, In the step S4, the warning judgment of the settlement trend prediction module further comprises the following analysis process: extracting the current temperature measurement value T recorded by the temperature compensation unit inst the difference ΔT from the reference temperature value T ref when the absolute value of ΔT exceeds 5°C for three consecutive periods, activating the temperature anomaly correction mode In the temperature anomaly correction mode, a temperature compensation verification rule is added to the condition of triggering the third-level early warning signal: when the vertical displacement data changes by more than 5 mm in three consecutive periods, ΔD is calculated synchronously c , and if the ratio K>0.3, the early warning signal is downgraded to the observation-level warning measured . ​ In determining whether the vertical displacement change gradient of three adjacent sensor nodes exceeds 2 mm / m, the standard deviation σ of ΔT at the three nodes is calculated synchronously T When σ T > 2°C, the gradient threshold is adjusted from 2 mm / m to 2.5 mm / m; The curvature radius of the sedimentation distribution curve is calculated by introducing a temperature cumulative influence factor where ΔT i is the ΔT value of each hour within 24 hours before the current calculation period, n is the sampling number, and the design allowable value is multiplied by a correction coefficient [1+0.02×(C T -10)]. The warning signal sent in the step S5 is additionally provided with a temperature association identification code, when the absolute value of ΔT exceeds 10℃, a temperature anomaly mark is embedded in the warning signal, the mark triggers the mobile terminal of the terminal device to start a temperature calibration mode, and the temperature calibration mode forcibly calls the ground temperature monitoring data in the latest 3 hours to cross-verify the platinum resistance temperature sensor data.

8. The coastal protection barrier settlement monitoring method of claim 1, wherein, In the step S2, the synchronous triggering unit further comprises a dynamic acquisition mode switching function, and the running mode is as follows: The data acquisition module continuously monitors the change rate of the vertical displacement data, and when the vertical displacement change of any sensor node exceeds 3 mm for two consecutive periods, the acquisition mode is automatically switched to an emergency acquisition mode; In the emergency acquisition mode, the synchronous triggering unit shortens the acquisition period to 10 minutes, and preferentially transmits the data of all nodes within 50 meters upstream and downstream of the sensor node; The data buffer unit starts a block storage mechanism in the emergency acquisition mode, marks the preferentially transmitted data as a high-priority data packet, and transmits the high-priority data packet in real time through the 4G network channel, and the remaining data is retained in the local cache; In the step S4, the settlement tendency prediction module starts a fast prediction algorithm when receiving a high-priority data packet, the fast prediction algorithm uses an exponential smoothing method to extrapolate data of the last 6 periods, and when the vertical displacement acceleration of the extrapolation result exceeds 0.05 mm / min2, a secondary early warning signal is forcibly triggered; The Beidou short message transmission mode of the communication base station is switched to a continuous sending state in the emergency collection mode, and the continuous sending interval is set to 5 minutes.

9. The coastal protection barrier embankment settlement monitoring method of claim 1, wherein, In the step S5, the early warning signal processing of the terminal device further includes the following linkage control logic: When the computer terminal of the dam management center receives the first-level early warning signal, the emergency control interface is automatically activated, the emergency control interface is connected with the PLC controller of the dam drainage gate, the gate opening adjustment instruction is sent to the PLC controller, and the drainage gate opening adjustment amount AG=min(0.5*AS c , 100mm), wherein AS c is the vertical displacement amount of the sensor node triggering the early warning. When the mobile terminal of the field maintenance personnel receives a secondary or higher early warning signal, the positioning and navigation function is automatically started, the coordinate data of the Beidou positioning module is called, and the optimal path from the current position to the early warning sensor node is generated; The special-level early warning signal triggered in the step S4 is synchronously transmitted to the maritime supervision platform through the communication base station, and the maritime supervision platform generates a ship navigation prohibited area electronic fence with a radius of 500 meters according to the early warning position coordinates; All linkage control operations record the operation time stamp and execution parameters, and are returned to the data processing center through the data collection module for closed-loop verification.

10. The coastal protection barrier embankment settlement monitoring method of claim 1, wherein, In the step S5, the early warning server sends the early warning signal to the terminal device through the communication base station using a 4G network and a Beidou short message dual-channel transmission mode, and the following dynamic selection mechanism is used: A signal quality monitoring unit is arranged in the communication base station, the signal quality monitoring unit collects the signal strength value RSSI of the 4G network and the time delay data of the Beidou short message in real time, and when the RSSI is lower than -90 dBm and the time delay data exceeds 5 seconds, the forced switching logic is activated; The forced switching logic performs the following operations: the data packets containing red identification codes in the early warning signal are preferentially transmitted using the Beidou short message, and the remaining data packets are cached locally and transmitted after the 4G signal is restored; One hour before the daily tidal level peak period, the communication base station automatically starts the dual-channel redundant transmission mode, in which the third-level and above early warning signals are sent twice through the 4G network and the Beidou short message at the same time, and the interval between the two transmissions is set to 30 seconds; The data buffer unit of the data collection module records the actual transmission success number of each data packet, and when it is detected that the same data packet fails in both channels, the field sound and light alarm is automatically triggered and a communication failure log is generated; The special-level early warning signal generated by the settlement tendency prediction module in the step S4 is attached with an emergency retransmission mark during transmission, and when the communication base station receives a data packet with the emergency retransmission mark, if the first transmission fails, it is repeatedly sent with an interval of 2 minutes until an acknowledgement receipt is received from the terminal device; When the mobile terminal of the terminal device receives data in the Beidou short message mode, it automatically compares the data integrity of the 4G channel and the Beidou channel, and when it is detected that the data loss rate exceeds 20%, it actively initiates a retransmission request to the communication base station.

Images