Coastal protection dam settlement monitoring method

Abstract

The invention discloses a coastal protection dam settlement monitoring method, and belongs to the technical field of hydraulic engineering safety monitoring. The method comprises the steps that a longitudinal monitoring section is arranged on the slope surface of the back sea side of a dam, a three-measuring-line fiber grating sensor array is arranged, and vertical displacement and horizontal dip angle data are periodically collected through a synchronous triggering unit; establishing a vertical displacement-horizontal dip angle joint analysis model by using a multi-source data fusion module, and eliminating tide level interference through a Kalman filtering algorithm to generate a settlement distribution curve; and triggering third-level to first-level early warning signals based on the grading early warning rule, and transmitting the signals to the terminal equipment. The problems that a traditional monitoring method cannot effectively separate tidal interference, the real-time performance is poor, and multi-dimensional data collaborative analysis is insufficient are solved, high-precision settlement monitoring, complex environment anti-interference and rapid emergency response are achieved through multi-measuring-line sensor deployment, dynamic filtering optimization and a graded early warning mechanism, and the method is suitable for large-scale popularization and application. And the reliability and timeliness of dam safety monitoring are obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of safety monitoring of hydraulic engineering, and particularly to a method for monitoring the settlement of coastal protection dikes. Background Art

[0002] As a key facility to resist tidal erosion and storm surge attacks, the structural stability of coastal protection dikes is directly related to the safety of people's lives and property in coastal areas. Due to the long-term influence of complex environments such as tidal cyclic loads, soil infiltration, and wave impacts, the dikes are prone to problems such as uneven settlement, basal scouring, and local instability.

[0003] Traditional settlement monitoring mainly relies on manual inspections, single-point settlement pile measurements, or fixed sensor networks, with the following limitations: Insufficient data timeliness: The manual inspection period is long (usually once a month), making it difficult to capture sudden settlements in a timely manner; although single-point sensors can continuously monitor, they lack the ability to perform collaborative analysis of multiple measurement lines and cannot comprehensively reflect the overall deformation trend of the dike.

[0004] Difficulty in separating environmental interference: The periodic water level change of the tide will cause elastic deformation of the dike, which is easily misjudged after being superimposed with the real settlement. Existing technologies mostly adopt static filtering or simple threshold methods, and it is difficult to dynamically eliminate interference factors such as tidal levels and temperatures.

[0005] Single warning mechanism: Traditional methods usually trigger warnings based on a single parameter (such as displacement), lacking collaborative judgment of multi-dimensional indicators such as settlement gradient and curvature change, resulting in a relatively high false alarm or missed alarm rate.

[0006] Limited communication reliability: Coastal areas are often affected by harsh environments such as typhoons and salt fog. Wired transmission is easily interrupted, and the bandwidth and stability of wireless communication (such as GPRS) are insufficient, making it difficult to ensure the real-time transmission of key warning information.

[0007] In existing improvement schemes, for example, fiber Bragg grating sensors are used to improve the monitoring accuracy, but their layout methods mostly focus on a single line at the top of the dike and it is difficult to capture the characteristics of layered settlement; some studies attempt to introduce the Kalman filtering algorithm, but do not perform dynamic parameter adjustment by combining multi-source data such as tidal levels and inclinations, and the filtering effect is limited. Therefore, there is an urgent need for a dike settlement monitoring method that can integrate multi-dimensional data, dynamically resist interference, and adapt to complex communication environments. Summary of the Invention

[0008] The present invention overcomes the problems of traditional monitoring methods that cannot effectively separate tidal interference, poor real-time performance, and insufficient collaborative analysis of multi-dimensional data. Through multi-measurement line sensor deployment, dynamic filtering optimization, and a hierarchical warning mechanism, it realizes high-precision settlement monitoring, anti-interference in complex environments, and rapid emergency response, significantly improving the reliability and timeliness of dike safety monitoring.

[0009] To achieve the above object, the present invention adopts the following solutions: A method for monitoring the settlement of a coastal protection dike, comprising the following steps: S1: Set longitudinal monitoring sections on the seaward slope of the dike body. The longitudinal monitoring sections extend along the dike axis. The longitudinal monitoring sections include a crest measurement line arranged on the dike crest, a toe measurement line arranged at the dike toe, and an intermediate measurement line arranged on the intermediate platform of the dike body. Fiber Bragg grating sensor arrays are respectively arranged at intervals on the crest measurement line, the toe measurement line and the intermediate measurement line. The fiber Bragg grating sensor array includes a plurality of sensor nodes arranged at intervals along the dike axis direction. Each sensor node includes a displacement meter arranged in the vertical direction and an inclination sensor arranged in the horizontal direction; S2: Set a data acquisition module in the monitoring control station. The data acquisition module is respectively connected to the three groups of fiber Bragg grating sensor arrays of the longitudinal monitoring section through armored optical cables. The data acquisition module includes a synchronous trigger unit and a data buffer unit. The synchronous trigger unit periodically sends synchronous acquisition instructions to all sensor nodes. The data buffer unit receives and stores the vertical displacement data measured by the displacement meter and the horizontal inclination data measured by the inclination sensor; S3: Configure a multi-source data fusion module 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 tidal level changes through the Kalman filtering algorithm, and generates a settlement distribution curve in the dike axis direction; S4: Deploy a settlement trend prediction module in 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 change amount of the vertical displacement data of any sensor node exceeds the change amount threshold D for N consecutive periods, a third-level early warning signal is triggered. When the vertical displacement change gradient of three adjacent sensor nodes exceeds the gradient threshold B, a second-level early warning signal is triggered. When the numerical value of the radius of curvature of the settlement distribution curve is less than the design allowable value C, a first-level early warning signal is triggered; S5: The early warning server sends an early warning signal to the terminal device through the communication base station. The communication base station adopts a dual-channel transmission mode of 4G network and Beidou short message. The terminal device includes a computer terminal of the dike management center and a mobile terminal of on-site maintenance personnel.

[0010] Preferably, in the step S3, establishing the vertical displacement-horizontal inclination joint analysis model includes the following steps: Divide the vertical displacement data into discrete segments at 10-minute intervals according to the time series, calculate the standard deviation and the linear fitting slope of the displacement change rate for each discrete segment, and generate a vertical displacement characteristic parameter sequence; Performing a fast Fourier transform on the horizontal tilt angle data, extracting the energy proportion of the 0.1-0.5 Hz frequency band as the tilt angle dynamic feature, and determining it as abnormal vibration interference when the energy proportion exceeds 60%; Establish a correlation coefficient matrix between the vertical displacement characteristic parameter sequence and the inclination dynamic characteristic quantity. 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 inclination coupling. The measured value of the elastic modulus of the main structure of the dam is introduced into the Kalman filter algorithm. The process noise covariance matrix of the filter is adjusted according to the product of the measured value of the elastic modulus and the dynamic characteristic value of the inclination angle. The adjustment formula is: , 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 inclination angle; The vertical displacement data of the section marked as affected by tilt coupling are input into the adjusted Kalman filter to generate the settlement distribution curve including the influence of structural stiffness, while the data points corresponding to abnormal vibration interference are eliminated.

[0011] Preferably, in step S3, after generating the settlement distribution curve including the influence of the structural stiffness, the following steps are further performed: 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 sequence of the top, middle and foot measuring lines in the same time period is extracted, and the displacement coordination index η between the three measuring lines is calculated. , where σ top , σ mid , σ foot are the standard deviations of the vertical displacement change rates of the top, middle and toe lines, respectively; Performing nonlinear fitting on the displacement coordination index η and 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 stratified settlement trend; 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 value ΔR=γ×(η-η0) is automatically added to the curvature radius value of the corresponding embankment section in the settlement distribution curve, where η0 is the benchmark coordination index during embankment construction acceptance, and γ is the inverse of the cross-sectional width of the embankment body. In step S4, the following judgment rule is added to the conditions for triggering the first-level warning signal: 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 warning signal; When generating the settlement distribution curve, the multi-source data fusion module synchronously outputs a three-dimensional deformation cloud map containing the layered settlement trend identifier, and the distribution range of η values is distinguished by color gradient in the three-dimensional deformation cloud map.

[0012] Preferably, in step S3, after generating the three-dimensional deformation cloud map containing the layered settlement trend identifier, the following steps are further performed: Select the cross-section where the sensor node with the lowest η value is located among the crest measurement line, the middle measurement line, and the toe measurement line. Use the mobile detection vehicle to carry the ground-penetrating radar to perform profile scanning along this cross-section. The ground-penetrating radar uses an 800 MHz antenna to collect dielectric constant data at an interval of 0.2 meters, and simultaneously use the static cone penetration equipment to measure the cone tip resistance value 1 meter away from the sensor node; Input the dielectric constant data and the cone tip resistance value into the multi-source data fusion module to establish a dielectric constant-cone tip resistance joint calibration model, and the model outputs the soil density correction coefficient μ of the main body of the dam, , where ε r is the relative value of the dielectric constant, and q c is the cone tip resistance value; Adjust the calculation formula of the displacement coordination index η according to the μ value, adjust it to η' = η×(1 + 0.2μ), and substitute the adjusted displacement coordination index η' back into the η-E relationship surface to determine the layered settlement trend; In step S2, the synchronous trigger 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 step S5, when sending the warning signal, add the soil density correction information. When μ < 1.2, add a red identification code to the warning signal; when 1.2 ≤ μ ≤ 1.5, add a yellow identification code; when μ > 1.5, add a green identification code; 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 sending 3 times per hour, the yellow identification code corresponds to sending 1 time per hour, and the green identification code corresponds to sending 1 time per 3 hours.

[0013] Preferably, in step S3, the vertical displacement-horizontal inclination joint analysis model further includes the following processing procedures: Access the tide level data transmitted in real time by the tide level monitoring station at the data processing center. The tide level monitoring station is located on the sea side of the main body of the dam and 10 meters away from the toe measurement line. The tide level monitoring station includes a pressure type water level gauge and a Beidou positioning module. The pressure type water level gauge measures the tide level elevation at a period of 5 minutes, and the Beidou positioning module records the longitude and latitude coordinates of the tide level monitoring station; The multi-source data fusion module synchronizes the timestamps of the tide level elevation and the vertical displacement data, extracts the periodic displacement component caused by the tide level change through the Kalman filtering algorithm, the extraction frequency range of the periodic displacement component is 0.5 - 2 times per hour, and this component is removed from the vertical displacement data; According to the correlation analysis result of the horizontal inclination data and the tide level elevation, dynamically adjust the weight coefficient of the inclination sensor in the joint analysis model, and the adjustment formula of the weight coefficient α is: , where ΔH is the difference in tide level elevation between two adjacent times, and k is the proportionality factor corresponding to the soil permeability coefficient of the main body of the dam; In the step S4, when the settlement trend prediction module triggers a secondary warning signal, the following operations are synchronously executed: Extract the tide level phase data of the tide level monitoring station within 24 hours before the current moment, calculate the difference between the actual tide level phase and the tide level phase in the same period of history, and when the difference exceeds 30 degrees and the vertical displacement change gradient exceeds 2 mm / m, upgrade the secondary warning signal to a primary warning signal; After generating the settlement distribution curve, according to the time delay correlation between the tide level elevation data and the vertical displacement data, correct the displacement compensation amount of the sensor nodes within 200 meters from the tide level monitoring station in the settlement distribution curve, and the displacement compensation amount is , where H inst is the current tide level elevation, H mean is the average tide level elevation of the current month, and β is the tangent value of the inclination angle of the sea-facing slope of the main body of the dam.

[0014] Preferably, in the step S1, the sensor nodes of the fiber Bragg grating sensor array further have the following structural features: A temperature compensation unit is arranged between the displacement gauge and the inclination sensor of each sensor node. The temperature compensation unit includes a platinum resistance temperature sensor closely attached to the surface of the main body of the dam, and the platinum resistance temperature sensor measures the contact temperature at the sensor node with a resolution of 0.1 °C; The output end of the temperature compensation unit is connected to the data acquisition module, and the data cache unit synchronously stores the temperature measurement data. The temperature measurement data has the same timestamp as the vertical displacement data and the horizontal inclination data; In the step S3, when the multi-source data fusion module establishes a vertical displacement-horizontal inclination joint analysis model, input the temperature measurement data into the displacement compensation model, and the displacement compensation model corrects the vertical displacement data in real time according to the temperature-displacement relationship curve. The calculation formula of the vertical displacement correction amount is: , where α T is the temperature sensitivity coefficient of the displacement gauge, T inst is the current temperature measurement value, and T ref is the reference temperature value during installation and commissioning; Before processing the vertical displacement data by the Kalman filtering algorithm, the corrected ΔD c is subjected to cumulant deduction, and the deduction period is consistent with the 1-hour period of the synchronous acquisition instruction.

[0015] Preferably, in step S4, the early warning judgment of the settlement trend prediction module further includes the following analysis process: Extract the current temperature measurement value T recorded by the temperature compensation unit inst and the difference ΔT from the reference temperature value T ref . When the absolute value of ΔT exceeds 5°C for three consecutive periods, the temperature anomaly correction mode is activated; In the temperature anomaly correction mode, a temperature compensation verification rule is added to the condition for triggering a level-three early warning signal: when the change amount of the vertical displacement data exceeds 5 mm for three consecutive periods, synchronously calculate ΔD c and the ratio measured of the measured displacement change amount ΔD . If the ratio K>0.3, the early warning signal is downgraded to an observation-level warning; When determining whether the vertical displacement change gradient of three adjacent sensor nodes exceeds 2 mm / m, synchronously calculate the standard deviation σ of ΔT at these three nodes T . When σ T >2°C, the gradient threshold is adjusted from 2 mm / m to 2.5 mm / m; When calculating the numerical value of the curvature radius of the settlement distribution curve, a temperature cumulative influence factor is introduced, where ΔT i is the ΔT value per hour within the previous 24 hours before the current calculation period, n is the number of sampling times, and the design allowable value is multiplied by the correction coefficient [1 + 0.02×(C T -10)]; The early warning signal sent in step S5 is appended with a temperature correlation identification code. When the absolute value of ΔT exceeds 10°C, a temperature anomaly mark is embedded in the early warning signal, and this mark 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 nearest 3 hours for cross-verification with the data of the platinum resistance temperature sensor.

[0016] Preferably, in step S2, the synchronous trigger unit further includes a dynamic acquisition mode switching function, and its operation mode is as follows: The data acquisition module continuously monitors the change rate of the vertical displacement data. When the change amount of the vertical displacement of any sensor node exceeds 3 mm for two consecutive periods, it automatically switches to the emergency acquisition mode; In the emergency acquisition mode, the synchronous trigger unit shortens the acquisition period to 10 minutes and preferentially transmits the data of all nodes within 50 meters upstream and downstream of this sensor node; In the emergency acquisition mode, the data cache unit starts the block storage mechanism, marks the data to be preferentially transmitted as high-priority data packets, and the high-priority data packets are transmitted in real time through the 4G network channel, while the remaining data is retained in the local cache. In step S4, when the settlement trend prediction module receives a high-priority data packet, it starts the fast prediction algorithm. The fast prediction algorithm uses the exponential smoothing method to extrapolate and calculate the data of the last 6 cycles. When the vertical displacement acceleration of the extrapolation result exceeds 0.05 mm / min², a secondary warning signal is forcibly triggered. In the emergency acquisition mode, the Beidou short message transmission mode of the communication base station is switched to the continuous sending state, and the continuous sending interval is set to 5 minutes.

[0017] Preferably, in step S5, the warning signal processing of the terminal device further includes the following linkage control logic: When the computer terminal of the dam management center receives a primary warning signal, it automatically activates the emergency control interface. The emergency control interface is connected to the PLC controller of the dam drainage gate, and sends a gate opening adjustment instruction to the PLC controller. The drainage gate opening adjustment amount ΔG = min(0.5×ΔS c , 100mm), where ΔS c is the vertical displacement of the sensor node that triggers the warning. When the mobile terminal of the on-site maintenance personnel receives a secondary or higher warning signal, it automatically starts 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 position to the warning sensor node. 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 for a ship no-sailing area with a radius of 500 meters according to the warning position coordinates. All linkage control operations record the operation timestamp and execution parameters, and are transmitted back to the data processing center through the data acquisition module for closed-loop verification.

[0018] Preferably, in step S5, the warning server sends warning signals to the terminal device through the communication base station using a dual-channel transmission mode of 4G network and Beidou short message, and uses the following dynamic selection mechanism: A signal quality monitoring unit is set in the communication base station. The signal quality monitoring unit real-time collects the signal strength value RSSI of the 4G network and the delay data of the Beidou short message. When RSSI is lower than -90dBm and the delay data exceeds 5 seconds, the forced switching logic is activated. The forced switching logic performs the following operations: the data packets containing the red identification code in the warning signal are preferentially transmitted using Beidou short messages, and the remaining data packets are cached locally and retransmitted after the 4G signal is restored; One hour before the daily peak tide period, the communication base station automatically activates the dual-channel redundant transmission mode. In this redundant transmission mode, warning signals of level 3 and above are sent twice simultaneously via the 4G network and Beidou short messages, with an interval of 30 seconds between the two transmissions. The data cache unit of the data acquisition module records the actual number of successful transmissions of each data packet. When it is detected that the same data packet fails in both dual-channel transmissions, the on-site sound and light alarm is automatically triggered and a communication failure log is generated. The special warning signal generated by the settlement trend prediction module in step S4 is attached with an emergency retransmission mark when it is transmitted. When the communication base station receives a data packet with the emergency retransmission mark, if the first transmission fails, it will be repeatedly sent at intervals of 2 minutes until a confirmation receipt is received from the terminal device; 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 it detects that the data missing rate exceeds 20%, it actively initiates a retransmission request to the communication base station.

[0019] The present invention has at least the following beneficial effects: (1) by deploying three-line sensors and integrating multi-source data, the coordinated monitoring of vertical settlement, horizontal deformation and environmental interference of the dam is realized, and the comprehensiveness and real-time performance of settlement analysis are significantly improved; (2) based on the optimization of stiffness parameters and the 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 tide interference and using 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 status to overall structural safety is realized, and the risk classification response capability is improved; (5) by using emergency acquisition mode and dual-channel redundant transmission, the integrity of monitoring data and the reliable delivery of early warning instructions in extreme environments are guaranteed, and the robustness of the system is enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a principle flow chart of the coastal protection dam settlement monitoring method provided by the present invention. DETAILED DESCRIPTION

[0021] The present invention will be described in further detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0022] like Figure 1 As shown, the coastal protection dam settlement monitoring method provided by the present invention includes the following steps: S1: Set up longitudinal monitoring sections on the seaward slope of the dam body. The longitudinal monitoring sections extend along the dam axis. The longitudinal monitoring sections include a crest measuring line arranged on the dam crest, a toe measuring line arranged at the toe of the dam, and an intermediate measuring line arranged on the intermediate platform of the dam body. Fiber Bragg grating sensor arrays are respectively arranged at intervals on the crest measuring line, the toe measuring line, and the intermediate measuring line. The fiber Bragg grating sensor array includes a plurality of sensor nodes arranged at intervals along the dam axis. Each sensor node includes a displacement meter arranged in the vertical direction and an inclination sensor arranged in the horizontal direction.

[0023] Due to long-term seawater erosion and soil infiltration on the seaward side of the dam, it is a sensitive area for settlement and deformation. Selecting this area to set up monitoring sections can preferentially capture the starting points of potential failures. In the structure of the longitudinal monitoring section, the crest measuring line is located in the middle of the dam crest and is used to monitor the compression deformation of the dam crest caused by self-weight and vehicle loads; the intermediate platform measuring line is arranged on the structural platform at half of the dam body height, reflecting the consolidation settlement characteristics of the middle part of the dam body; the toe measuring line is arranged close to the toe line to monitor the scour settlement of the foundation caused by tidal level changes.

[0024] The arrangement method of the fiber Bragg grating sensor array is to arrange sensor nodes every 20 meters along the dam axis (a conventional interval selection, which can also be adjusted to 10 - 30 meters according to the dam length); the sensor types include vertical displacement meters, which use vibrating wire displacement meters to measure vertical displacement through the change of the steel wire frequency (range ±50mm, resolution 0.01mm); it also includes inclination sensors, which measure the horizontal inclination based on MEMS technology (range ±15°, accuracy 0.01°).

[0025] The crest measuring line focuses on the overall settlement, the intermediate measuring line monitors the layered settlement trend, and the toe measuring line captures the foundation deformation, forming a three-dimensional monitoring network; the 20-meter interval of sensor nodes can balance the monitoring accuracy and engineering cost, and avoid missing local deformations.

[0026] S2: Set up a data acquisition module in the monitoring control station. The data acquisition module is respectively connected to the three groups of fiber Bragg grating sensor arrays of the longitudinal monitoring section through armored optical cables. The data acquisition module includes a synchronous trigger unit and a data buffer unit. The synchronous trigger unit periodically sends synchronous acquisition instructions to all sensor nodes, and the data buffer unit receives and stores the vertical displacement data measured by the displacement meter and the horizontal inclination data measured by the inclination sensor.

[0027] A double-layer stainless steel armored optical cable is used to connect the sensor and the monitoring station, with a tensile strength ≥ 2000N to prevent seawater corrosion and mechanical damage. The synchronous trigger unit uses periodic commands, with a 1-hour synchronous acquisition period (a conventional selection, which can also be adjusted to 2 hours according to the tidal cycle); a synchronous mechanism is also adopted to send pulse signals through the optical cable to ensure that the timestamp error of all node data < 10ms. The data buffer unit adopts a circular buffer design, with a storage capacity ≥ 72 hours of raw data; vertical displacement data (16-bit precision), tilt angle data (floating-point type), and timestamp (UTC format). The trigger unit sends a synchronous pulse every hour, and the sensor node completes the displacement and tilt angle measurement within 10ms; the data is transmitted to the buffer unit through the optical cable, and data packets are generated according to "survey line number + node position" to avoid data confusion.

[0028] S3: Configure a multi-source data fusion module in the data processing center. The multi-source data fusion module receives the vertical displacement data and horizontal tilt angle data transmitted by the data acquisition module, and establishes a vertical displacement-horizontal tilt angle joint analysis model. The model eliminates the periodic displacement interference caused by tidal level changes through the Kalman filter algorithm, and generates a settlement distribution curve in the direction of the dam axis.

[0029] The joint analysis model architecture includes: Input layer: vertical displacement data (time series), horizontal tilt angle data (spatial distribution); Processing layer: The Kalman filter algorithm eliminates tidal periodic interference, and the spatial interpolation algorithm generates a continuous settlement curve (such as the Kriging interpolation method); The output layer outputs the settlement distribution curve (the horizontal axis is the distance along the dam axis, and the vertical axis is the settlement amount).

[0030] After the data fusion module receives the transmitted raw data, it first aligns the timestamps; corrects the direction deviation of the displacement measurement through the tilt angle data (for example, displacement compensation is triggered when the tilt angle > 1°); after eliminating the tidal periodic interference by the Kalman filter, a net settlement distribution curve is generated.

[0031] S4: Deploy a settlement trend prediction module on the warning server. The settlement trend prediction module connects to the multi-source data fusion module and obtains the settlement distribution curve. When the change amount of the vertical displacement data of any sensor node exceeds the change amount threshold D for N consecutive periods, a third-level warning signal is triggered. When the vertical displacement change gradient of three adjacent sensor nodes exceeds the gradient threshold B, a second-level warning signal is triggered. When the numerical value of the radius of curvature of the settlement distribution curve is less than the design allowable value C, a first-level warning signal is triggered.

[0032] Adopt a warning rule with three - level differentiation. Among them, the three - level warning is for single - point continuous settlement (for example, when N = 3 and D = 5mm, it means that the cumulative settlement in 3 periods > 15mm), indicating local soil softening; the two - level warning is for the excessive settlement gradient of adjacent nodes (for example, B = 2mm / m), indicating the formation of a potential slip surface; the condition for the first - level warning: the radius of curvature < the design value (for example, C = 500m, this value is referred to the industry requirements of the dam structure design, that is, when the radius of curvature R < 500m or is close to this condition, it is triggered), indicating the risk of overall instability.

[0033] The prediction module architecture conducts real - time analysis, including updating the settlement curve and warning status every 1 hour, making historical comparisons, and calling the construction - period acceptance data as the benchmark reference. When the settlement amounts of a certain node reach 5.2mm, 5.5mm, and 5.8mm continuously for 3 hours, the three - level warning is triggered; if the settlement gradient of the nodes within 50 meters upstream and downstream of it reaches 2.3mm / m, it is upgraded to the two - level warning; when the radius of curvature of the settlement curve of this section drops to 480m (the design allowable value is 500m), the first - level warning is triggered.

[0034] S5: The warning server sends warning signals to the terminal devices through the communication base station. The communication base station adopts a dual - channel transmission mode of 4G network and Beidou short message. The terminal devices include the computer terminals of the dam management center and the mobile terminals of on - site maintenance personnel.

[0035] The communication base station is configured with a 4G network for regular data transmission (bandwidth ≥ 10Mbps), and Beidou short message can be used as an emergency channel (the single - transmission capacity is 78 bytes, supporting positioning and feedback). Professional monitoring software is deployed on the computer terminals of the terminal devices to display the settlement curve and warning level; the mobile terminals receive warning information and navigation paths through the APP. After the warning server generates a warning signal, it first sends it to the management center through the 4G network; if the 4G signal is interrupted (such as in typhoon weather), it automatically switches to Beidou short message to transmit coordinates and warning codes; after the on - site personnel's mobile terminals receive the warning, they are automatically navigated to the target node (error < 5m).

[0036] This method can achieve the following through the above steps: Precise monitoring ability: Through the layout of three measurement lines and the fusion of multiple sensors, synchronous capture of vertical settlement and horizontal deformation is realized, and the missed detection rate is reduced to less than 5%; Anti - interference optimization: Kalman filtering effectively separates tidal - cycle interference, and the signal - to - noise ratio of settlement data is increased to a very high level; Hierarchical response efficiency: The three - level warning mechanism shortens the emergency response time to a very short time, and the recognition accuracy of major dangerous situations is very high; Communication robustness: The dual - channel transmission ensures that the data reach rate is above 99.5% in extreme environments.

[0037] In another technical solution, in step S3, establishing a vertical displacement - horizontal inclination joint analysis model includes the following steps: Dividing the vertical displacement data into discrete segments with 10-minute intervals according to the time series, calculating the standard deviation of the displacement change rate and the linear fitting slope for each discrete segment, and generating a vertical displacement characteristic parameter sequence; Performing a fast Fourier transform on the horizontal tilt angle data, extracting the energy proportion of the 0.1-0.5 Hz frequency band as the tilt angle dynamic feature, and determining it as abnormal vibration interference when the energy proportion exceeds 60%; Establish a correlation coefficient matrix between the vertical displacement characteristic parameter sequence and the inclination dynamic characteristic quantity. 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 inclination coupling. The measured value of the elastic modulus of the main structure of the dam is introduced into the Kalman filter algorithm. The process noise covariance matrix of the filter is adjusted according to the product of the measured value of the elastic modulus and the dynamic characteristic value of the inclination angle. The adjustment formula is: , 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 inclination angle; The vertical displacement data of the section marked as affected by tilt coupling are input into the adjusted Kalman filter to generate the settlement distribution curve including the influence of structural stiffness, while the data points corresponding to abnormal vibration interference are eliminated.

[0038] Vertical displacement data are divided into discrete segments with fixed time intervals (e.g., 10-minute segments). Characteristic parameters reflecting settlement dynamics are extracted by calculating the standard deviation of the displacement change rate and the slope of the linear fit within each segment. The standard deviation quantifies the intensity of settlement fluctuations, while the linear slope indicates the direction of settlement trends. Horizontal inclination data undergoes a Fast Fourier Transform (FFT) and analyzes the energy contribution of the 0.1-0.5 Hz frequency band (the conventional vibration monitoring frequency band). When the energy in this frequency band exceeds 60% of the total energy (an adjustable threshold of 50%-70% depending on geological conditions), it is identified as abnormal vibration interference (e.g., non-settlement deformation caused by mechanical construction or wave impact). By comparing the temporal correlation between vertical displacement characteristics and inclination dynamic characteristics (when the absolute value of the correlation coefficient exceeds 0.7), periods affected by inclination coupling are identified to avoid misinterpreting structural torsion as settlement.

[0039] Introduce the measured value of the dam elastic modulus (obtained by drilling and sampling during the construction period) into the Kalman filtering algorithm, and dynamically adjust the filtering parameters in combination with the dynamic characteristic quantity of the inclination angle. The elastic modulus reflects the anti-deformation ability of the material. When the elastic modulus is relatively high (for example, greater than 50 MPa), the adjustment range of the noise covariance matrix in the filtering process increases, enhancing the sensitivity to the deformation of rigid structures. The adjusted filter can distinguish the displacement differences caused by soil creep (low frequency) and structural stiffness changes (medium and high frequencies), thereby generating a more accurate settlement distribution curve. During implementation, the elastic modulus data is regularly updated through the Internet of Things to ensure that the model parameters are consistent with the actual working conditions.

[0040] Input the data marked as the inclination angle coupling action section into the optimized filter, and synchronously eliminate the abnormal vibration data points with an energy ratio exceeding the standard (such as eliminating the time period with an energy ratio > 60%). When outputting the settlement curve, add a stiffness identifier to the area affected by the structural stiffness (such as the concrete slope protection section) to prompt key attention to the material fatigue risk. For example, in the monitoring of sea dikes, by this method, the elimination ratio of the inclination angle interference data caused by wave impact is reduced from 25% to 8%, and the settlement calculation error is reduced to ±1.2 mm.

[0041] Precisely identify non-settlement interference through frequency band energy analysis, significantly improve the recognition ability of non-settlement interference signals, enhance data validity; improve the monitoring sensitivity to material property changes by fusing structural stiffness parameters; optimize the output of the settlement curve to provide a more explicit guiding basis for targeted maintenance.

[0042] In step S3 described above, after generating the settlement distribution curve including the influence of structural stiffness, the following steps are also executed: According to the time periods corresponding to the correlation coefficients with absolute values greater than 0.7 in the correlation coefficient matrix, extract the vertical displacement characteristic parameter sequences of the top of the dam measurement line, the middle measurement line, and the toe of the dam measurement line in the same time period, and calculate the displacement coordination index η between the three measurement lines. , where σ top , σ mid , σ foot are the standard deviations of the vertical displacement change rates of the top of the dam measurement line, the middle measurement line, and the toe of the dam measurement line respectively. Perform a non-linear fitting on the displacement coordination index η and the measured value of the elastic modulus to establish an η-E relationship surface. When the curvature of the η-E relationship surface exceeds the preset threshold, it is determined that there is a trend of layered settlement in the main body of the dam. Add a stiffness anomaly detection unit to the settlement trend prediction module. The stiffness anomaly detection unit receives the η-E relationship surface data. When detecting a trend of layered settlement, it automatically adds a correction amount ΔR = γ×(η - η0) to the curvature radius value of the corresponding dam section in the settlement distribution curve, where η0 is the reference coordination index during the dam construction acceptance, and γ is the reciprocal of the width of the cross-section of the main body of the dam. In step S4, the following determination rules are added to the conditions for triggering a first-level warning signal: when the value of the corrected radius of curvature is simultaneously less than the design allowable value and the curvature of the η-E relationship surface exceeds a preset threshold, the warning signal is upgraded to a special-level warning signal; When the multi-source data fusion module generates the settlement distribution curve, it synchronously outputs a three-dimensional deformation cloud map containing the layered settlement trend identifier, and the distribution range of η values is distinguished by color gradient in the three-dimensional deformation cloud map.

[0043] Based on the correlation coefficient matrix, high-correlation time periods (correlation coefficient > 0.7) are screened, and the vertical displacement characteristic parameters of the three measuring lines at the top, middle, and bottom of the embankment are extracted at the same time period. By calculating the ratio relationship of the standard deviations of the displacement change rates of the three measuring lines (such as the difference degree between the top and middle measuring lines, and the difference degree between the middle and bottom measuring lines), the displacement coordination index η is defined. The closer the η value is to 1, the better the deformation coordination of the three measuring lines; when the η value is lower than 0.5 (adjustable threshold), it indicates the risk of layered settlement. For example, in the monitoring of soft foundation embankments, if the η value drops suddenly from 0.8 to 0.3, it corresponds to the accurate depth of the soil stripping layer detected by subsequent ground penetrating radar.

[0044] The η value and the elastic modulus E are non-linearly fitted (such as quadratic surface fitting) to establish the η-E relationship surface. When the curvature of the surface exceeds a preset threshold (such as curvature > 0.05 / m), it is determined that there is a trend of layered settlement. During implementation, the curvature threshold is trained through historical data to ensure the adaptability to the geological conditions of different embankment sections. The stiffness anomaly detection unit adds a correction amount to the radius of curvature of the settlement curve according to the real-time η-E data (for example, correcting the original radius of curvature of 500m to 480m), making the warning judgment closer to the actual structural state.

[0045] The generated three-dimensional deformation cloud map visually shows the distribution of the layered settlement trend through color gradient (such as red represents η < 0.4, and green represents η > 0.8). When the corrected radius of curvature is simultaneously less than the design value (such as 500m) and the η-E curvature exceeds the standard, the first-level warning is upgraded to a special-level warning. In a certain engineering case, this method warns of a potential landslide 12 hours in advance, enabling the emergency repair team to reinforce the embankment foot in time.

[0046] Through the collaborative analysis of multiple measuring lines, the recognition accuracy of the layered settlement trend is significantly improved, and the ability to predict potential risks is enhanced; the η-E surface model realizes the coupled warning of stiffness and deformation, effectively reducing misjudgments caused by changes in structural stiffness and improving the reliability of warnings; the deformation characteristics are visually displayed through three-dimensional visualization technology, greatly improving the decision-making support efficiency.

[0047] In step S3, after generating the three-dimensional deformation cloud map containing the layered settlement trend identifier, the following steps are also executed: Select the cross-section where the sensor node with the lowest η value is located among the crest measurement line, the middle measurement line, and the toe measurement line. Use a mobile detection vehicle equipped with a ground penetrating radar to perform profile scanning along this cross-section. The ground penetrating radar uses an 800 MHz antenna to collect dielectric constant data at 0.2-meter intervals, and simultaneously use a static cone penetration test device to measure the cone tip resistance value at a distance of 1 meter from the sensor node. Input the dielectric constant data and the cone tip resistance value into the multi-source data fusion module to establish a dielectric constant-cone tip resistance joint calibration model. The model outputs the soil density correction coefficient μ of the main body of the dam. , where ε r is the relative value of the dielectric constant, and q c is the cone tip resistance value. Adjust the calculation formula of the displacement coordination index η according to the μ value, adjust it to η' = η × (1 + 0.2μ), and substitute the adjusted displacement coordination index η' back into the η-E relationship surface to determine the layered settlement trend. In step S2, the synchronous trigger 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 step S5, when sending the warning signal, add the soil density correction information. When μ < 1.2, add a red identification code to the warning signal; when 1.2 ≤ μ ≤ 1.5, add a yellow identification code; when μ > 1.5, add a green identification code. 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 sending 3 times per hour, the yellow identification code corresponds to sending 1 time per hour, and the green identification code corresponds to sending 1 time per 3 hours.

[0048] Deploy a mobile detection vehicle at the cross-section with the lowest η value (usually the weakest area), equipped with a ground penetrating radar (such as an 800 MHz antenna) to scan at high density (0.2-meter intervals) to obtain the dielectric constant distribution, reflecting the change of soil moisture content. Simultaneously use a static cone penetrometer (penetration speed 2 cm / s) to measure the cone tip resistance and obtain the soil shear strength data. Correlate the dielectric constant and the cone tip resistance through a joint calibration model (for example, when ε r × q c > 10, it is determined as a dense area), and output the soil density correction coefficient μ to quantify the local soil stability.

[0049] The data collection cycle is dynamically adjusted based on the μ value: shortened to 30 minutes when μ is less than 1.2 (loose soil), and extended to two hours when μ is greater than 1.8 (compact soil). Warning signals are assigned red, yellow, and green identification codes (for example, red corresponds to μ less than 1.2, requiring immediate action). Communication base stations adjust the Beidou message transmission frequency based on these identification codes (red three times per hour, green once every three hours). This approach has tripled data collection frequency in high-risk areas of reclamation dams and dikes, reducing communication resource consumption by 40%.

[0050] After receiving the alert, the mobile terminal automatically navigates to the target node (with a positioning error of less than 3 meters) and displays the real-time μ value to assist with on-site assessment. For example, if a yellow alert (μ = 1.4) is detected during on-site inspection, no emergency measures are required, thus avoiding an overreaction. This density data is also fed back to the data processing center to update the regional weight parameters of the η-E relationship model.

[0051] On-site verification of geological parameters significantly improves the precision of on-site verification of soil conditions and enhances local risk assessment capabilities; dynamically optimizes data collection 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 coordination.

[0052] In another technical solution, in step S3, the vertical displacement-horizontal inclination angle joint analysis model further includes the following processing steps: The data processing center receives tide data transmitted in real time from a tide monitoring station. The tide monitoring station is located on the seaward side of the main dam body and 10 meters from the embankment foot measurement line. The tide monitoring station includes a pressure water level gauge and a Beidou positioning module. The pressure water level gauge measures the tide elevation in a 5-minute cycle, and the Beidou positioning module records the latitude and longitude coordinates of the tide monitoring station. The multi-source data fusion module synchronizes the timestamps of the tidal elevation and the vertical displacement data, and extracts the periodic displacement component caused by the tidal level change through the Kalman filter algorithm. The extraction frequency range of the periodic displacement component is 0.5-2 times / hour, and this component is removed from the vertical displacement data. According to the correlation analysis results between horizontal tilt angle data and tide elevation, the weight coefficient of the tilt angle sensor in the joint analysis model is dynamically adjusted. The weight coefficient α adjustment formula is: , where ΔH is the difference between two adjacent tidal elevations, and k is the proportional factor corresponding to the permeability coefficient of the main soil of the dam; In step S4, when the secondary warning signal is triggered, the settlement trend prediction module simultaneously performs the following operations: Extract the tidal phase data of the tidal monitoring station within 24 hours before the current time, calculate the difference between the actual tidal phase and the historical tidal phase during the same period, and upgrade the Level 2 warning signal to a Level 1 warning signal when the difference exceeds 30 degrees and the vertical displacement change gradient exceeds 2 mm / m; After generating the settlement distribution curve, according to the time-delay correlation between the tide level elevation data and the vertical displacement data, correct the displacement compensation amount of the sensor nodes within 200 meters from the tide level monitoring station in the settlement distribution curve. The displacement compensation amount is , where H inst is the current tide level elevation, H mean is the average monthly tide level elevation, and β is the tangent value of the slope inclination angle on the sea-facing side of the main body of the dam.

[0053] A tide level monitoring station is arranged 10 meters away from the toe line on the sea-facing side of the dam (the location is selected to take into account both the directly affected area of the tide and the equipment safety). The monitoring 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 the water pressure (range 0 - 10 meters, accuracy ±1 cm), and collects data every 5 minutes (which can be adjusted to 2 - 10 minutes to adapt to the tide change speed). The Beidou module records the longitude and latitude coordinates of the monitoring station (positioning accuracy ±0.5 meters) to ensure the accurate association of the tide level data with the dam position. The data is accessed in real-time to the data processing center through wireless transmission and aligned with the time stamp of the vertical displacement data (the error is controlled within ±10 seconds). For example, in the estuary dam, the synchronization deviation between the tide level data and the displacement data can be shortened from the original 30 seconds to 8 seconds, significantly improving the reliability of the correlation analysis.

[0054] The multi-source data fusion module separates the periodic influence of the tide level through the Kalman filtering algorithm. The algorithm identifies the displacement component caused by the tide level change (frequency range 0.5 - 2 times / hour, corresponding to the characteristics of semi-diurnal tide to diurnal tide), and eliminates it from the original displacement data. At the same time, the weight coefficient of the tilt sensor is dynamically adjusted according to the tide level elevation difference (such as the change amount of the tide level between two adjacent times ΔH = 0.8 meters). The weight coefficient is related to the soil permeability coefficient of the dam (for example, sandy soil k = 0.3, clay k = 0.1). The stronger the soil permeability, the higher the influence weight of the tide level change on the tilt. In implementation, when ΔH exceeds 1 meter (such as during a storm surge), the weight coefficient automatically increases to 0.9 to enhance the sensitivity of the tilt data to the structural tilt.

[0055] When a secondary warning is triggered (the settlement gradient of three adjacent nodes exceeds the standard), the tidal phase difference within 24 hours is analyzed synchronously. If the deviation between the actual tidal phase and the historical same period exceeds 30 degrees (an adjustable threshold, such as 20 - 40 degrees), and the settlement gradient continues to exceed the standard, the warning will be upgraded to the first level. 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 the first-level warning in advance to avoid the expansion of cracks at the dike crest. In addition, for sensor nodes within 200 meters of the tidal level monitoring station (areas significantly affected by tidal scour), the displacement compensation amount is calculated based on the difference between the current tidal level and the monthly average value. In the calculation of the compensation amount, the tangent value β of the slope inclination angle is obtained from the design drawing (for example, a slope ratio of 1:3 corresponds to β = 0.33), and after correction, the short-term fluctuation of the tidal level is eliminated to interfere with the long-term settlement trend.

[0056] The effective fusion of tidal level data eliminates the periodic interference of tides and improves the authenticity of settlement data analysis; the dynamic weight adjustment mechanism dynamically adjusts the weights of sensors, enhancing the monitoring adaptability under complex working conditions; the phase difference correlation warning upgrade significantly improves the collaborative warning ability of compound disasters through the phase correlation mechanism, and strives for a critical time window for emergency repair decisions.

[0057] In another technical solution, in the step S1, the sensor nodes of the fiber Bragg grating sensor array further include 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 includes a platinum resistance temperature sensor closely attached to the surface of the dike body, and the platinum resistance temperature sensor measures the contact temperature at the sensor node with a resolution of 0.1 °C; The output end of the temperature compensation unit is connected to the data acquisition module, and the data cache unit synchronously stores the temperature measurement data. The temperature measurement data has the same time stamp as the vertical displacement data and the horizontal inclination data; In the step S3, when the multi-source data fusion module establishes a vertical displacement-horizontal inclination joint analysis model, the temperature measurement data is input into the displacement compensation model, and the displacement compensation model corrects the vertical displacement data in real time according to the temperature-displacement relationship curve. The calculation formula for the vertical displacement correction amount is: , where α T is the temperature sensitivity coefficient of the displacement meter, T inst is the current temperature measurement value, T ref is the reference temperature value during installation and commissioning; Before processing the vertical displacement data by the Kalman filter algorithm, the cumulative amount of the corrected ΔD c is deducted first, and the deduction period is consistent with the 1-hour period of the synchronous acquisition instruction.

[0058] Integrate a temperature compensation unit between the displacement meter and the inclinometer of each sensor node. This unit uses a platinum resistance temperature sensor (PT100 type, temperature measurement range -20°C to 80°C) installed closely to the dam surface to measure the contact temperature in real time with a resolution of 0.1°C. The temperature data is transmitted to the data acquisition module through an independent channel and shares the same timestamp (error < 10 ms) with the displacement and inclination data to ensure multi-parameter synchronous analysis. For example, in the monitoring of cold-region seawalls, the temperature sensor detects a daily temperature difference of up to 25°C, and the synchronous displacement data fluctuates by up to 1.8 mm, verifying the necessity of temperature compensation. The sensor node housing is designed with waterproof and sealed protection (protection level IP68) to prevent seawater penetration from causing temperature measurement distortion.

[0059] The multi-source data fusion module calls the 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 calculation of the correction amount depends on the temperature sensitivity coefficient of the displacement meter (for example, for a certain model of displacement meter, α T = 0.02 mm / °C). When the deviation between the on-site temperature and the installation reference temperature (usually the annual average temperature) exceeds ±5°C, the compensation calculation is automatically triggered. During implementation, the data processing center uniformly deducts the cumulative correction amount (for example, a cumulative correction of 1.0 mm in 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 ±1.5 mm to ±0.3 mm.

[0060] Before the Kalman filter algorithm formally processes the data, it first preprocesses the vertically displaced data after temperature correction. The preprocessing includes: eliminating the linear drift caused by temperature (such as deducting the cumulative amount of 0.02 mm / °C per hour); marking the data during the period of sudden temperature change (such as ΔT > 8°C within 1 hour) for subsequent analysis and reference. In the dams in tropical regions, it can successfully identify false settlement alarms caused by a sudden drop in temperature (ΔT = 12°C in 2 hours) due to heavy rain and avoid false triggering of the third-level warning.

[0061] Temperature compensation significantly improves the displacement measurement accuracy in a temperature-changing environment and reduces the influence of environmental interference; the preprocessing mechanism effectively reduces false alarms caused by sudden temperature changes and enhances the system stability; the multi-parameter timestamp alignment enhances the reliability of data correlation verification through multi-parameter synchronous analysis.

[0062] In step S4 described above, the early warning judgment of the settlement trend prediction module further includes the following analysis process: Extract the current temperature measurement value T recorded by the temperature compensation unit inst and the difference ΔT from the reference temperature value T ref . When the absolute value of ΔT exceeds 5°C for 3 consecutive cycles, activate the temperature anomaly correction mode; In the temperature anomaly correction mode, add a temperature compensation verification rule to the conditions for triggering a level-three warning signal: when the change amount of the vertical displacement data exceeds 5 mm in three consecutive cycles, synchronously calculate ΔD c and the measured displacement change amount ΔD measured ratio . If the ratio K > 0.3, downgrade the warning signal to an observation-level alarm; When determining whether the vertical displacement change gradient of three adjacent sensor nodes exceeds 2 mm / m, synchronously calculate the standard deviation σ of ΔT at these three nodes T . When σ T > 2°C, adjust the gradient threshold from 2 mm / m to 2.5 mm / m; When calculating the numerical value of the curvature radius of the settlement distribution curve, introduce a temperature cumulative influence factor , where ΔT i is the ΔT value per hour within the previous 24 hours before the current calculation period, n is the number of sampling times, and multiply the design allowable value by the correction factor [1 + 0.02×(C T - 10)]; The warning signal sent in step S5 is appended with a temperature correlation identification code. When the absolute value of ΔT exceeds 10°C, embed a temperature anomaly mark in the warning signal, and this mark 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 nearest 3 hours for cross-verification with the data of the platinum resistance temperature sensor.

[0063] When the absolute value of the temperature deviation ΔT in three consecutive acquisition cycles (which can be set to 1 - 5 cycles) exceeds 5°C (an adjustable threshold, set to 3 - 8°C according to the regional climate), the system activates the temperature anomaly correction mode. In this mode, triggering a level-three warning requires additional verification of the temperature correction amount ΔD c and the ratio K of the measured displacement change amount. For example, when ΔD c accounts for more than 30% (i.e., K > 0.3), it indicates that the current displacement change is mainly caused by temperature expansion, and the warning is downgraded to an observation-level alarm (only recorded without triggering an emergency response). Through this rule, a certain sea dike can reduce the number of false alarms per day from more than ten times to 2 times or less during the high-temperature period in summer.

[0064] When calculating the settlement gradient of three adjacent nodes, synchronously analyze the standard deviation σ of the temperature deviation in this area T . When σ T > 2°C (which can be set to 1.5 - 3°C), it indicates that the uneven distribution of the temperature field may amplify the local settlement difference. At this time, relax the original gradient threshold of 2 mm / m to 2.5 mm / m (the adjustment amplitude can increase with the increase of σ_T). For example, if σ at three nodes during monitoring T= 2.8℃, the system automatically adopts a threshold of 2.6 mm / m to avoid misjudgment caused by uneven sunlight irradiation.

[0065] The temperature cumulative effect compensation and terminal response mechanism introduces the temperature cumulative influence factor C T (average ΔT value in 24 hours) to correct the design allowable value. When C T > 10℃ (such as in continuous high-temperature weather), the amplification factor of the design allowable value increases by 0.02 / ℃ to adapt to the thermal expansion effect of the material. After the warning signal is embedded with the temperature anomaly mark, the mobile terminal automatically starts the calibration mode: call the data of the nearby ground temperature monitoring station (if any) for cross-verification. Display the temperature influence ratio (e.g., the temperature contribution degree in the current displacement is 45%) for on-site judgment. In a cold wave event, this mechanism can help engineers quickly distinguish between real settlement (temperature contribution degree < 20%) and frost heave deformation (temperature contribution degree > 60%).

[0066] The temperature-related warning rules significantly optimize the warning logic in temperature anomaly scenarios, reducing unnecessary emergency responses; the dynamic gradient threshold mechanism improves the monitoring adaptability of complex temperature fields; through temperature contribution degree analysis, it enhances the scientific nature of on-site disposal decisions.

[0067] In another technical solution, in the step S2, the synchronous trigger unit further includes a dynamic acquisition mode switching function, and its operation mode is as follows: The data acquisition module continuously monitors the change rate of the vertical displacement data. When the vertical displacement change amount of any sensor node exceeds 3 mm in two consecutive cycles, it automatically switches to the emergency acquisition mode; In the emergency acquisition mode, the synchronous trigger unit shortens the acquisition cycle to 10 minutes and preferentially transmits the data of all nodes within 50 meters upstream and downstream of this sensor node; The data cache unit starts a block storage mechanism in the emergency acquisition mode, marks the preferentially transmitted data as high-priority data packets, and the high-priority data packets are transmitted in real time through the 4G network channel, and the remaining data is retained in the local cache; In the step S4, when the settlement trend prediction module receives the high-priority data packet, it starts a fast prediction algorithm. The fast prediction algorithm uses the exponential smoothing method to extrapolate and calculate the data of the last 6 cycles. When the vertical displacement acceleration of the extrapolation result exceeds 0.05 mm / min², it forcibly triggers a secondary warning signal; The Beidou short message transmission mode of the communication base station switches to the continuous sending state in the emergency acquisition mode, and the continuous sending interval is set to 5 minutes.

[0068] When the vertical displacement change of any sensor node exceeds 3 mm (adjustable threshold, set to 2 - 5 mm according to the geological risk level) in two consecutive acquisition periods (which can be set to 1 - 3 periods), the synchronization trigger unit automatically switches to the emergency acquisition mode. In this mode, the acquisition period is shortened from the regular 1 hour to 10 minutes (adjustable range is 5 - 15 minutes), and the data of all sensors within 50 meters upstream and downstream of the abnormal node is preferentially transmitted (covering the potential impact area). For example, at the initial stage of a piping danger, the system increases the acquisition frequency by 6 times within 2 hours, and timely captures the accelerated settlement trends of 5 surrounding nodes. The data cache unit starts the block storage mechanism, and transmits the high-priority data packets (marked in red) in real time through the 4G network, and the regular data is temporarily stored locally (retained for up to 24 hours) to ensure that key information is not lost.

[0069] After receiving the high-priority data packet, the settlement trend prediction module immediately starts the extrapolation calculation using the exponential smoothing method. The algorithm calculates the vertical displacement acceleration based on the last 6 periods (i.e., 6 times of 10-minute data within 1 hour) (for example, 0.06 mm / min² exceeds the threshold of 0.05). If the predicted value continues to exceed the standard, a secondary warning is forcibly triggered (skipping the regular analysis process). At the same time, the Beidou short message transmission switches to the continuous sending mode (interval of 5 minutes, adjustable to 3 - 10 minutes) to ensure that at least 12 key data transmissions are made per hour under extreme network interruptions. During a typhoon, this mechanism can shorten the warning response time from the regular 45 minutes to 18 minutes.

[0070] When transmitting high-priority data packets, the communication base station automatically allocates 80% of the bandwidth resources (adjustable ratio) to ensure their priority passage. All emergency operation records (such as mode switching time, extrapolation parameters) are attached to the data packet metadata for post-event analysis. When the communication is restored, the system automatically re-transmits the locally cached data, and optimizes the algorithm parameters by comparing the predicted value with the actual value (for example, adjusting the exponential smoothing coefficient).

[0071] The emergency mode significantly improves the data capture density at the initial stage of the danger situation, shortens the emergency response window, and buys precious time for decision-making; it speeds up the identification of major risks through the forced warning mechanism; and the dynamic bandwidth allocation optimizes the communication resource configuration to ensure the timeliness of key data transmission.

[0072] In another technical solution, in step S5, the warning signal processing of the terminal device further includes the following linkage control logic: When the computer terminal of the dam management center receives a first-level warning signal, it automatically activates the emergency control interface. The emergency control interface is connected to the PLC controller of the dam drainage gate, and sends a gate opening adjustment instruction to the PLC controller. The drainage gate opening adjustment amount ΔG = min(0.5×ΔS c , 100mm), where ΔS cThe vertical displacement of the sensor node that triggers the early warning; When the mobile terminal of the on-site maintenance personnel receives an early warning signal of level two or above, the positioning and navigation function is automatically activated. The positioning and navigation function calls the coordinate data of the Beidou positioning module to generate the optimal path from the current position to the early warning sensor node; The special level early warning signal triggered in step S4 is synchronously transmitted to the maritime supervision platform through the communication base station, notifying the maritime supervision platform to generate an electronic fence for the ship no-go area with a radius of 500 meters according to the early warning position coordinates; All linkage control operations record the operation timestamp and execution parameters, and are transmitted back to the data processing center through the data acquisition module for closed-loop verification.

[0073] After receiving the level one early warning signal, the computer terminal of the dam management center immediately activates the emergency control interface. This interface connects to the PLC controller of the drainage gate through the Modbus protocol and sends an opening adjustment instruction. The adjustment amount ΔG is calculated based on the node displacement amount ΔS that triggers the early warning (for example, when ΔS = 20mm, ΔG = 10mm), and ΔG = min(0.5×ΔS c , 100mm) means that the adjustment amount takes the smaller value of 0.5×ΔS c and 100mm, and a maximum adjustment limit of 100mm is set to prevent excessive drainage. At the same time, the mobile terminal (equipped with a customized APP) of the on-site maintenance personnel automatically activates the Beidou navigation function, generates the optimal path based on the early warning node coordinates (avoiding the landslide area, with a path planning error < 3 meters). In a certain actual combat drill, this function shortened the arrival time of the repair team from an average of 25 minutes to 12 minutes.

[0074] The special level early warning signal is synchronously transmitted to the maritime supervision platform through the communication base station. After the platform analyzes the early warning coordinates, it automatically generates an electronic no-go area with a radius of 500 meters (adjustable to 300 - 800 meters). The no-go instruction is broadcast to the surrounding ships through the AIS system, and the channel indicator lights are linked to switch to red warning. For example, after the landslide early warning is triggered, the electronic fence can successfully intercept fishing boats that stray into the dangerous area and avoid secondary accidents.

[0075] All linkage operations (such as gate opening, navigation path, no-go instruction) record the accurate timestamp and execution parameters (such as the actual action delay of the gate < 2 seconds). The data acquisition module monitors the operation effect in real time (such as the displacement change rate of the corresponding dam section after drainage), and transmits the feedback data back to the processing center for effect verification. If the operation fails to meet the expectation (such as the displacement does not slow down within 1 hour), the system automatically upgrades the early warning level or switches the disposal plan.

[0076] The automatic adjustment of the floodgate greatly improves the automation level of emergency control and accelerates the execution efficiency of emergency operations; the electronic fence effectively reduces the risk of secondary disasters and enhances the collaborative disposal ability of multiple departments; the closed-loop verification mechanism improves the effectiveness and traceability of disposal measures.

[0077] In step S5, the early warning server sends early warning signals to the terminal device through the communication base station using a dual-channel transmission mode of 4G network and Beidou short message, and uses the following dynamic selection mechanism: A signal quality monitoring unit is set in the communication base station. The signal quality monitoring unit real-time collects the signal strength value RSSI of the 4G network and the delay data of the Beidou short message. When the RSSI is lower than -90dBm and the delay data exceeds 5 seconds, the forced switching logic is activated; The forced switching logic performs the following operations: for the data packets containing red identification codes in the early warning signal, the Beidou short message is preferentially used for transmission, and the remaining data packets are cached locally and wait to be retransmitted after the 4G signal is restored; One hour before the daily high tide period, the communication base station automatically starts the dual-channel redundant transmission mode. In the redundant transmission mode, the early warning signals of level three and above 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 cache unit of the data acquisition module records the actual transmission success times of each data packet. When it is detected that the same data packet fails in both channels of the dual-channel transmission, the on-site sound and light alarm is automatically triggered and a communication failure log is generated; In step S4, the special early warning signal generated by the settlement trend prediction module is attached with an emergency retransmission mark during transmission. When the communication base station receives a data packet with an emergency retransmission mark, if the first transmission fails, it will be resent at intervals of 2 minutes until an acknowledgment receipt from the terminal device is received; 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. When it is detected that the data loss rate exceeds 20%, a retransmission request is actively sent to the communication base station.

[0078] The communication base station is built-in with a signal quality monitoring unit, which real-time collects the 4G signal strength (RSSI) and the Beidou message delay. When the RSSI is lower than -90dBm (can be set to -85 to -100dBm to adapt to different devices) and the delay exceeds 5 seconds (can be adjusted to 3 - 8 seconds), the forced switching logic is triggered: the data with red identification code (such as high-risk early warning with μ < 1.2) is preferentially sent through Beidou (single packet size ≤ 78 bytes), and the remaining data is temporarily stored locally (reserved for up to 12 hours). One hour before the daily high tide (can be set in advance), the base station starts the dual-channel redundant transmission mode, and repeats the transmission of early warning signals of level three and above twice (interval 30 seconds) to ensure at least one successful reception.

[0079] The data cache unit records the actual number of times each data packet is sent and its success status. If the same data packet fails on both channels (e.g., three consecutive transmissions with no response), an on-site audible and visual alarm (volume ≥ 90 decibels, flashing frequency 2 Hz) is immediately triggered, and a fault log with location information is generated. For special warning packets, an emergency retransmission flag (highest priority) is added. If the initial transmission fails, the packet is sent repeatedly at 2-minute intervals (adjustable to 1-5 minutes) until an acknowledgment (ACK signal) is received from the terminal. In the event of a fiber optic cable outage, this mechanism enables the special warning packet to be successfully delivered within 15 minutes, while conventional 4G channel recovery takes up to 6 hours.

[0080] When receiving data in Beidou mode, the mobile terminal automatically compares the data integrity of the 4G and Beidou channels (for example, verifying packet sequence number continuity). If the missing packet rate exceeds 20% (adjustable to 15-25%), the terminal proactively initiates a retransmission request (with a list of missing packet IDs) to the base station. This retransmission request is sent via a Beidou message (occupying a dedicated channel), and the base station retransmits data based on priority. The terminal also activates an offline cache function (storage capacity ≥ 48 hours of data) to ensure critical information can be retrieved during network outages.

[0081] Intelligent channel switching significantly enhances the robustness of communications in extreme environments and ensures the accessibility of critical information; improves the reliability of transmission of special warnings through an intelligent retransmission mechanism; and optimizes the terminal's autonomous verification function to reduce the interference of data missing on decision-making.

[0082] It should be noted that although the steps are described above in a specific order, this does not necessarily mean that the steps must be performed in this specific order. In fact, some of these steps can be performed concurrently or even in a different order, as long as the required functions can be achieved. The number of devices and processing scales described here are intended to simplify the description of the present invention. Applications, modifications, and variations of the present invention will be apparent to those skilled in the art.

[0083] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the description and implementation methods. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A method for monitoring the settlement of a coastal protection dike, characterized in that, The following steps are involved: S1: A longitudinal monitoring section is set up on the seaward slope of the main body of the dam. The longitudinal monitoring section extends along the axis of the dam. The longitudinal monitoring section includes a crest measurement line arranged at the crest of the dam, a foot measurement line arranged at the foot of the dam, and a middle measurement line arranged at the middle platform of the dam. Fiber Bragg grating sensor arrays are set up at intervals on the crest measurement line, the foot measurement line, and the middle measurement line. The fiber Bragg grating sensor array includes multiple sensor nodes arranged at intervals along the axis of the dam. Each sensor node includes a displacement meter arranged in the vertical direction and an inclination sensor arranged in the horizontal direction. S2: A data acquisition module is set up in the monitoring and control station. The data acquisition module is connected to the three groups of fiber grating sensor arrays in 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 instructions to all sensor nodes. The data buffer unit receives and stores the vertical displacement data measured by the displacement meter and the horizontal tilt data measured by the tilt 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 horizontal inclination data transmitted by the data acquisition module and establishes a vertical displacement-horizontal inclination joint analysis model. The model uses the Kalman filter algorithm to eliminate the periodic displacement interference caused by tidal level changes and generates a settlement distribution curve along the dam axis. 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 exceeds the change threshold D for N consecutive cycles, a third-level early warning signal is triggered. When the vertical displacement change gradient of three adjacent sensor nodes exceeds the gradient threshold B, a second-level early warning signal is triggered. When the curvature radius of the settlement distribution curve is less than the design allowable value C, a first-level early warning signal is triggered. S5: The early warning server sends early warning signals to the terminal devices through the communication base station. The communication base station adopts the dual-channel transmission mode of 4G network and Beidou short message. The terminal devices include the computer terminals of the dam management center and the mobile terminals of the on-site maintenance personnel.

2. The method for monitoring the settlement of a coastal protection dike according to claim 1, wherein In step S3, establishing a vertical displacement-horizontal inclination joint analysis model includes the following steps: Dividing the vertical displacement data into discrete segments with 10-minute intervals according to the time series, calculating the standard deviation of the displacement change rate and the linear fitting slope for each discrete segment, and generating a vertical displacement characteristic parameter sequence; Performing a fast Fourier transform on the horizontal tilt angle data, extracting the energy proportion of the 0.1-0.5 Hz frequency band as the tilt angle dynamic feature, and determining it as abnormal vibration interference when the energy proportion exceeds 60%; Establish a correlation coefficient matrix between the vertical displacement characteristic parameter sequence and the inclination dynamic characteristic quantity. 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 inclination coupling. Introduce the measured value of the elastic modulus of the dam body structure into the Kalman filtering algorithm, and adjust the process noise covariance matrix of the filter according to the product of the measured value of the elastic modulus and the dynamic characteristic quantity of the inclination angle. The adjustment formula is: , where Q is the adjusted process noise covariance matrix, Q0 is the original process noise covariance matrix, E is the measured value of the elastic modulus, and A is the dynamic characteristic quantity of the inclination angle; The vertical displacement data of the section marked as affected by tilt coupling are input into the adjusted Kalman filter to generate the settlement distribution curve including the influence of structural stiffness, while the data points corresponding to abnormal vibration interference are eliminated.

3. The settlement monitoring method of the coastal protection dike according to claim 2, wherein, In step S3, after generating the settlement distribution curve considering the influence of structural stiffness, the following steps are further executed: According to the time periods corresponding to the correlation coefficients with absolute values greater than 0.7 in the said correlation coefficient matrix, extract the vertical displacement characteristic parameter sequences of the top dike survey line, the middle survey line and the toe dike survey line in the same time period, and calculate the displacement coordination index η between the three survey lines, , where σ top , σ mid , σ foot are the standard deviations of the vertical displacement change rates of the top dike survey line, the middle survey line and the toe dike survey line respectively; Non-linearly fit the displacement coordination index η with the measured value of the elastic modulus to establish the η-E relationship surface. When the curvature of the η-E relationship surface exceeds a preset threshold, it is determined that there is a trend of layered settlement in the main body of the dam; Add a stiffness anomaly detection unit to the settlement trend prediction module. The stiffness anomaly detection unit receives the η-E relationship surface data. When a layered settlement trend is detected, it automatically adds a correction amount ΔR = γ×(η - η0) to the curvature radius value of the corresponding dam section in the settlement distribution curve, where η0 is the reference coordination index during the dam construction acceptance, and γ is the reciprocal of the cross-sectional width of the main body of the dam; In step S4, the following determination rule is added to the condition for triggering the first-level warning signal: When the corrected curvature radius value is simultaneously 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; When generating the settlement distribution curve, the multi-source data fusion module synchronously outputs a three-dimensional deformation cloud map containing the layered settlement trend identifier, and the distribution range of η values is distinguished by color gradient in the three-dimensional deformation cloud map.

4. The coastal protection dike settlement monitoring method according to claim 3, characterized in that, In step S3, after generating the three-dimensional deformation cloud map containing the layered settlement trend identifier, the following steps are further executed: Select the cross-section where the sensor node with the lowest η value is located among the top-of-dam survey line, middle survey line, and toe-of-dam survey line. Use a mobile detection vehicle to carry a ground penetrating radar to perform cross-sectional scanning along this cross-section. The ground penetrating radar uses an 800 MHz antenna to collect dielectric constant data at an interval of 0.2 meters, and simultaneously uses a static cone penetration test device to measure the cone tip resistance value 1 meter away from the sensor node; Input the dielectric constant data and the cone tip resistance value into the multi-source data fusion module to establish a dielectric constant-cone tip resistance joint calibration model, and the model outputs the compactness correction coefficient μ of the dam body soil mass. , where ε r is the relative value of the dielectric constant, and q c is the cone tip resistance value; Adjust the calculation formula of the displacement coordination index η according to the μ value, adjusted to η' = η×(1 + 0.2μ), and substitute the adjusted displacement coordination index η' back into the η-E relationship surface for layered settlement trend determination; In step S2, the synchronous trigger 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 step S5, soil density 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; when μ > 1.5, a green identification code is added; 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 transmissions per hour, the yellow identification code corresponds to 1 transmission per hour, and the green identification code corresponds to 1 transmission per 3 hours.

5. The method for monitoring the settlement of a coastal protection dike according to claim 1, characterized in that, In step S3, the vertical displacement-horizontal inclination joint analysis model further includes the following processing procedures: The tidal level data transmitted in real time by the tidal level monitoring station is accessed at the data processing center. The tidal level monitoring station is located on the sea-facing side of the main body of the dam and is 10 meters away from the toe measuring line. The tidal level monitoring station includes a pressure-type water level gauge and a Beidou positioning module. The pressure-type water level gauge measures the tidal level elevation at a cycle 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 synchronizes the timestamps of the tidal level elevation and the vertical displacement data, extracts the periodic displacement components caused by the tidal level change through the Kalman filter algorithm, and the extraction frequency range of the periodic displacement components is 0.5 - 2 times per hour, and removes this component from the vertical displacement data. According to the correlation analysis results of the horizontal inclination angle data and the tide level elevation, dynamically adjust the weight coefficient of the inclination sensor in the joint analysis model. The adjustment formula for the weight coefficient α is as follows: , where ΔH is the difference in tide level elevation between two adjacent times, and k is the proportionality factor corresponding to the soil permeability coefficient of the main body of the dam; In step S4, when the settlement trend prediction module triggers a secondary warning signal, the following operations are synchronously executed: Extract the tidal level phase data of the tidal level monitoring station within 24 hours before the current moment, calculate the difference between the actual tidal level phase and the tidal level phase in the same period of history. When the difference exceeds 30 degrees and the vertical displacement change gradient exceeds 2 mm / m, upgrade the secondary warning signal to a primary warning signal. After generating the settlement distribution curve, according to the time-delay correlation between the tide level elevation data and the vertical displacement data, correct the displacement compensation amount of the sensor nodes within 200 meters from the tide level monitoring station in the settlement distribution curve. The displacement compensation amount is , where H inst is the current tide level elevation, H mean is the monthly average tide level elevation, and β is the tangent value of the slope inclination angle on the sea-facing side of the main body of the dam.

6. The method for monitoring the settlement of a coastal protection dike according to claim 1, wherein In step S1, the sensor nodes of the fiber Bragg grating sensor array also include the following structural features: A temperature compensation unit is set between the displacement meter and the inclination sensor of each sensor node. The temperature compensation unit includes a platinum resistance temperature sensor closely attached to the surface of the main body of the dam. The platinum resistance temperature sensor measures the contact temperature at the sensor node with a resolution of 0.1 °C. The output end of the temperature compensation unit is connected to the data acquisition module. The data cache unit synchronously stores the temperature measurement data, and the temperature measurement data has the same timestamp as the vertical displacement data and the horizontal inclination data. In the step S3, when the multi-source data fusion module establishes a vertical displacement-horizontal inclination joint analysis model, the temperature measurement data is input into the displacement compensation model. The displacement compensation model corrects the vertical displacement data in real time according to the temperature-displacement relationship curve. The calculation formula for the vertical displacement correction amount is: , where α T is the temperature sensitivity coefficient of the displacement gauge, T inst is the current temperature measurement value, and T ref is the reference temperature value during installation and commissioning; Before processing the vertical displacement data, the Kalman filtering algorithm first performs cumulant deduction on the corrected ΔD c The deduction period is consistent with the 1-hour period of the synchronous acquisition instruction.

7. The coastal protection dike settlement monitoring method according to claim 6, characterized in that, In step S4, the warning judgment of the settlement trend prediction module also includes the following analysis process: Extract the current temperature measurement value T recorded by the temperature compensation unit inst and the reference temperature value T ref to obtain the difference ΔT. When the absolute value of ΔT exceeds 5°C for three consecutive cycles, activate the temperature anomaly correction mode; In the temperature anomaly correction mode, add a temperature compensation verification rule to the conditions for triggering a level-three warning signal: when the change in the vertical displacement data exceeds 5 mm for three consecutive cycles, synchronously calculate ΔD c and the measured displacement change ΔD measured ratio , if the ratio K > 0.3, downgrade the warning signal to an observation-level alarm; When determining whether the vertical displacement change gradient of three adjacent sensor nodes exceeds 2 mm / m, synchronously calculate the standard deviation σ of ΔT at these three nodes. T , when σ T > 2 °C, adjust the gradient threshold from 2 mm / m to 2.5 mm / m; When numerically calculating the curvature radius value of the settlement distribution curve, a temperature cumulative influence factor is introduced , where ΔT i is the ΔT value per hour within the previous 24 hours before the current calculation period, n is the number of sampling times, and the design allowable value is multiplied by the correction coefficient [1 + 0.02×(C T - 10)]; The warning signal sent in step S5 is attached with a temperature correlation identification code. When the absolute value of ΔT exceeds 10 °C, a temperature anomaly mark is embedded in the warning signal, and this mark 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 in the recent 3 hours for cross-verification with the data of the platinum resistance temperature sensor.

8. The method for monitoring the settlement of a coastal protection dike according to claim 1, characterized in that, In step S2, the synchronous trigger unit also includes a dynamic acquisition mode switching function, and its operation mode is as follows: The data acquisition module continuously monitors the change rate of the vertical displacement data. When the vertical displacement change amount of any sensor node in two consecutive cycles exceeds 3 mm, it automatically switches to the emergency acquisition mode. In the emergency acquisition mode, the synchronous trigger unit shortens the acquisition cycle to 10 minutes and preferentially transmits the data of all nodes within 50 meters upstream and downstream of this sensor node. The data cache unit starts a block storage mechanism in the emergency acquisition mode, marks the preferentially transmitted data as high-priority data packets, and the high-priority data packets are transmitted in real time through the 4G network channel, and the remaining data is retained in the local cache. In step S4, when the settlement trend prediction module receives a high-priority data packet, it starts the fast prediction algorithm. The fast prediction algorithm uses the exponential smoothing method to extrapolate the data of the last 6 cycles. When the vertical displacement acceleration of the extrapolation result exceeds 0.05 mm / min², a secondary warning signal is forcibly triggered. The Beidou short message transmission mode of the communication base station is switched to the continuous transmission state in the emergency acquisition mode, and the continuous transmission interval is set to 5 minutes.

9. The settlement monitoring method of the coastal protection dike according to claim 1, characterized in that In step S5, the warning signal processing of the terminal device further includes the following linkage control logic: When the computer terminal of the dam management center receives a first-level warning signal, it automatically activates the emergency control interface. The emergency control interface is connected to the PLC controller of the dam drainage gate and sends a gate opening adjustment instruction to the PLC controller. The drainage gate opening adjustment amount ΔG = min(0.5×ΔS c , 100mm), where ΔS c is the vertical displacement of the sensor node that triggers the warning; When the mobile terminal of the on-site maintenance personnel receives a warning signal of level 2 or above, the positioning and navigation function is automatically started. The positioning and navigation function calls the coordinate data of the Beidou positioning module to generate the optimal path from the current position to the warning sensor node. The special warning signal triggered in step S4 is synchronously transmitted to the maritime supervision platform through the communication base station, notifying the maritime supervision platform to generate an electronic fence for the ship navigation ban area with a radius of 500 meters according to the warning position coordinates. All linkage control operations record the operation timestamp and execution parameters, and are transmitted back to the data processing center through the data acquisition module for closed-loop verification.

10. The coastal protection levee settlement monitoring method according to claim 1, characterized in that In step S5, the warning server sends a warning signal to the terminal device through the communication base station using a dual-channel transmission mode of 4G network and Beidou short message, and uses the following dynamic selection mechanism: A signal quality monitoring unit is set in the communication base station. The signal quality monitoring unit real-time collects the signal strength value RSSI of the 4G network and the delay data of the Beidou short message. When RSSI is lower than -90dBm and the delay data exceeds 5 seconds, the forced switching logic is activated. The forced switching logic performs the following operations: The data packets containing the red identification code in the warning signal are preferentially transmitted using the Beidou short message, and the remaining data packets are cached locally and wait to be retransmitted after the 4G signal is restored. One hour before the daily high tide period, the communication base station automatically starts the dual-channel redundant transmission mode. In the redundant transmission mode, warning signals of level 3 and above 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 cache unit of the data acquisition module records the actual transmission success times of each data packet. When it is detected that the same data packet fails in both channels of the dual-channel transmission, the on-site sound and light alarm is automatically triggered and a communication failure log is generated. The special warning signal generated by the settlement trend prediction module in step S4 is attached with an emergency retransmission mark during transmission. When the communication base station receives a data packet with an emergency retransmission mark, if the first transmission fails, it will be repeatedly transmitted at intervals of 2 minutes until an acknowledgment receipt from the terminal device is received. 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. When the detected data loss rate exceeds 20%, a retransmission request is actively sent to the communication base station.

Images