Automatic monitoring device for displacement of vertical retaining structure and monitoring method thereof

Abstract

The invention discloses a vertical retaining structure displacement automatic monitoring device and a monitoring method thereof, and relates to the technical field of displacement automatic monitoring, and the device comprises a data acquisition unit which is used for configuring monitoring nodes, obtaining sensor data monitored at the top of a retaining wall, and preprocessing the sensor data to obtain effective inclination angle data; the displacement conversion unit is used for recognizing and optimizing the attitude angle by using a dynamic Kalman filtering algorithm, and calculating a displacement value of the vertical soil retaining structure by combining the vertical distance from the bottom to the top of the soil retaining wall; the monitoring and early warning unit is used for uploading the displacement value of the vertical retaining structure to a remote data platform, constructing a monitoring network, analyzing and predicting the deformation trend of the retaining wall and triggering early warning according to an analysis result; and the battery management unit is used for providing battery state monitoring and management. The displacement of the top of the retaining wall can be accurately sensed and monitored through the inclination angle data, and comprehensive and accurate monitoring is ensured by adopting an attitude solver and a dynamic Kalman filtering algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic displacement monitoring, and more specifically, to an automatic displacement monitoring device for a vertical retaining structure and a monitoring method thereof. Background Art

[0002] With the continuous development of urban construction, vertical retaining structures are widely used in civil engineering. To ensure the safety and stability of the structure, it is particularly important to monitor the displacement and deformation of the retaining wall in real time. However, the existing monitoring methods usually require complex manual measurements or rely on fixed sensors, which cannot meet the requirements of real-time, accurate, and long-term online monitoring. In addition, the traditional monitoring system lacks intelligent processing capabilities and cannot automatically analyze data and give early warnings, and the support for long-term online monitoring is also very limited.

[0003] For traditional manual measurement methods, a large amount of manpower and time are required. Moreover, due to the complex and changeable on-site conditions, the measurement results are often affected by environmental factors, resulting in poor data accuracy and consistency. In addition, manual measurement can only provide data at specific moments and cannot achieve continuous monitoring, which may cause potential problems to be ignored until they develop into serious safety hazards. For monitoring methods that rely on fixed sensors, although they reduce some manual dependence, they lack intelligent analysis capabilities and cannot automatically process data and give early warnings, and the support for long-term online monitoring is limited. Therefore, there is an urgent need for an automatic monitoring device that can achieve intelligent monitoring, real-time data transmission, remote management, and has high precision and long battery life.

[0004] Regarding the problems in the related art, no effective solution has been proposed yet. Summary of the Invention

[0005] In view of the problems in the related art, the present invention provides an automatic displacement monitoring device for a vertical retaining structure and a monitoring method thereof to overcome the above-mentioned technical problems existing in the related art.

[0006] To this end, the specific technical solutions adopted by the present invention are as follows:

[0007] In a first aspect, the present invention provides an automatic displacement monitoring device for a vertical retaining structure, including:

[0008] A data acquisition unit, configured to configure monitoring nodes, obtain sensor data monitored at the top of the retaining wall, and preprocess the sensor data to obtain effective inclination data;

[0009] A displacement conversion unit, configured to use the dynamic Kalman filtering algorithm to identify and optimize the attitude angle, and combine the vertical distance from the bottom to the top of the retaining wall to calculate the displacement value of the vertical retaining structure;

[0010] The monitoring and early warning unit is used to upload the displacement value of the vertical retaining structure to the remote data platform, construct a monitoring network, analyze and predict the deformation trend of the retaining wall, and trigger an early warning according to the analysis results.

[0011] Furthermore, the data acquisition unit includes:

[0012] The data acquisition module is used to acquire the sensor data monitored in real time by various types of sensors; the sensor data includes gyroscope data, accelerometer data, and geomagnetic sensor data;

[0013] The gyroscope preprocessing module is used to eliminate the static offset error of the gyroscope data based on an adaptive high-pass digital filter to obtain effective gyroscope data;

[0014] The accelerometer preprocessing module is used to perform time-domain smoothing on the accelerometer data based on a moving average filter, and separate the smoothed accelerometer data using a second-order Butterworth low-pass filter to obtain effective accelerometer data;

[0015] The geomagnetic data preprocessing module is used to eliminate noise from the geomagnetic sensor data based on a windowed median filter, identify the environmental magnetic field interference frequency band through fast Fourier transform frequency-domain analysis, and suppress the noise in the interference frequency band in combination with a notch filter to obtain effective geomagnetic sensor data;

[0016] The data integration module is used to integrate the processed effective gyroscope data, effective accelerometer data, and effective geomagnetic sensor data as the effective tilt angle data of the retaining wall.

[0017] Furthermore, identifying the environmental magnetic field interference frequency band through fast Fourier transform frequency-domain analysis includes:

[0018] Normalize the geomagnetic sensor data after noise elimination, and perform frequency-domain transformation on the normalized geomagnetic sensor data using the fast Fourier transform function to obtain the geomagnetic sensor data represented in the frequency domain;

[0019] According to the geomagnetic sensor data represented in the frequency domain, obtain the frequency corresponding to each complex number point, and calculate the amplitude spectrum of each frequency point;

[0020] Based on the amplitude spectrum of each frequency point, draw an amplitude spectrum diagram, and obtain the frequency corresponding to the peak point with a higher amplitude as the environmental magnetic field interference frequency band.

[0021] Furthermore, the displacement conversion unit includes:

[0022] The error accumulation processing module is used to obtain the tilt optimization angle at the previous moment, and combine it with the effective gyroscope data at the current moment to calculate the tilt optimization angle at the current moment;

[0023] An optimized angle correction module is used to correct the effective accelerometer data using the Kalman gain, combine with the tilt optimization calculation at the current moment, and obtain the optimized attitude angle through dynamic weight reduction calculation;

[0024] A displacement conversion module is used to obtain the vertical distance from the bottom to the top of the retaining wall measured by the laser rangefinder, and calculate the displacement value of the vertical retaining structure through conversion of the optimized attitude angle.

[0025] Furthermore, the calculation formula for the tilt optimization angle at the current moment is:

[0026]

[0027] In the formula, represents the tilt optimization angle at time k; represents the tilt optimization angle at the previous moment; ω k represents the gyroscope angular velocity corresponding to the effective gyroscope data; Δt represents the time interval;

[0028] The calculation formula for the optimized attitude angle is:

[0029]

[0030] In the formula, represents the optimized attitude angle; Z accel，k represents the accelerometer tilt angle at time k; K k represents the Kalman gain at time k;

[0031] The calculation formula for the displacement value of the vertical retaining structure is:

[0032] ΔL = h·tanθ k ;

[0033] In the formula, ΔL represents the displacement value of the vertical retaining structure; h represents the vertical distance from the bottom to the top of the retaining wall.

[0034] Furthermore, the monitoring and warning unit includes:

[0035] An edge communication and computing unit is used to converge the edge computing data to a remote data platform using a wireless communication protocol to achieve data aggregation of distributed monitoring nodes;

[0036] A digital modeling and display unit is used to construct a three-dimensional virtual model of the retaining wall, and map the deformation state of the physical entity in real time by inputting the displacement values of the vertical retaining objects at each monitoring node;

[0037] A spatio-temporal correlation prediction unit is used to predict the future deformation trend of the retaining wall and identify potential risk points based on historical and real-time data through spatio-temporal information fusion and mining;

[0038] An early warning decision support unit is used to set dynamic multi-level early warning strategies, trigger multi-level early warnings according to the analysis and prediction results, and generate emergency disposal and maintenance suggestions.

[0039] Furthermore, based on historical and real-time data, through spatio-temporal information fusion and mining, the future deformation trend of the retaining wall is predicted and potential risk points are identified, including:

[0040] Obtain the historical and real-time displacement values of the vertical retaining structure, sensor data and environmental data of the retaining wall, map all monitoring nodes to a unified spatio-temporal coordinate, and use the Pearson correlation coefficient to quantify the spatial correlation between monitoring nodes, screen strongly correlated neighbor nodes, and construct an adjacency matrix;

[0041] Based on the adjacency matrix, aggregate the features of adjacent monitoring nodes through a multi-layer graph convolutional network, normalize the adjacency matrix, and transfer the node features layer by layer to capture the synergy effect of local deformation of the retaining wall;

[0042] Input the spatially aggregated features into a bidirectional gated recurrent network, model the long-term dependence relationship of the displacement sequence along the time axis, introduce a spatio-temporal cross-attention mechanism, calculate the correlation weight between the spatial features and the temporal hidden state, and generate a coupled state vector that fuses spatio-temporal dynamics;

[0043] Based on the spatio-temporal fusion features, output the mean difference and variance of the displacement prediction value through a linear layer, construct a time-varying Gaussian analysis model, optimize the model parameters using the maximum likelihood loss function, and use the time-varying Gaussian analysis model to predict the mean and variance of the deformation amount at future moments;

[0044] According to the mean and variance prediction results, evaluate the deformation trend and potential risk points of the retaining wall, and introduce an online learning mechanism to optimize the model using the latest monitoring data.

[0045] Furthermore, the parameter prediction formula of the time-varying Gaussian analysis model is:

[0046] [[ID=2,7]]

[0047] In the formula, μ t represents the mean value of the predicted displacement value of the vertical retaining structure; W μ represents the linear layer weight of the mean value; c t represents the spatio-temporal fusion feature vector; b μ represents the mean bias term; σ t represents the variance of the predicted displacement value of the vertical retaining structure; W σ represents the linear layer weight of the variance; b σ represents the variance bias term;

[0048] The formula of the maximum likelihood loss function is:

[0049]

[0050] wherein, L represents the maximum likelihood loss function; y t represents the actual observed value of the displacement of the vertical retaining structure at time t; T represents the total length of the time series.

[0051] Furthermore, based on the mean and variance prediction results, evaluate the deformation trend and potential risk points of the retaining wall, and introduce an online learning mechanism to optimize the model using the latest monitoring data, including:

[0052] The future deformation of the retaining wall follows a Gaussian distribution. Based on the mean and variance prediction results, determine the probability interval in which the future deformation lies, and calibrate the risk points according to the deformation of each monitoring node;

[0053] Calculate the mutual information value between the environmental factors and the displacement value of the vertical retaining structure, dynamically adjust the input weights of the time-varying Gaussian distribution model, and construct an environment-threshold mapping table to match the current environmental adjustment standard;

[0054] Set an update period to regularly collect the latest monitoring data, calculate the loss value of the new data through the maximum likelihood loss function, and if the loss value is greater than the preset threshold, update the parameters of the time-varying Gaussian distribution model;

[0055] Map the mean prediction result to a three-dimensional virtual model to visually display the risk distribution of the retaining wall.

[0056] In a second aspect, the present invention also provides an automatic monitoring method for the displacement of a vertical retaining structure, and the monitoring method includes:

[0057] Configure distributed monitoring nodes at the top of the retaining wall, obtain the sensor data monitored at the top of the retaining wall, and preprocess the sensor data to obtain effective inclination data;

[0058] Based on the effective inclination data, use the dynamic Kalman filtering algorithm to identify and optimize the attitude angle, and combine the vertical distance from the bottom to the top of the retaining wall to calculate the displacement value of the vertical retaining structure;

[0059] Upload the displacement values of the vertical retaining structure calculated by all monitoring nodes to a remote data platform, construct a monitoring network, analyze and predict the deformation trend of the retaining wall, and trigger an alarm according to the analysis results.

[0060] The beneficial effects of the present invention are:

[0061] 1. The present invention can accurately sense and monitor the displacement at the top of the retaining wall through the tilt angle data, solving the defect that traditional monitoring means cannot obtain displacement data in real time and accurately. By adopting an attitude resolver and a dynamic Kalman filtering algorithm, and through a remote data platform for remote data transmission and control, automatic monitoring is achieved, and the monitoring data is more comprehensive and accurate, avoiding complex manual measurement. At the same time, edge computing is combined to perform real-time noise reduction and spatio-temporal alignment on the data, significantly improving the data reliability. The spatio-temporal prediction model based on graph convolutional network and recurrent neural network can dynamically analyze the overall deformation trend and local risk points of the retaining wall, quantify the uncertainty through probability prediction and trigger hierarchical early warning. It can support long-term continuous operation in an outdoor environment without power grid, and combined with a self-check mechanism and a remote maintenance interface, greatly reducing the frequency of manual inspections. The modular design takes into account the deployment convenience and system scalability, providing a high-precision, fully automatic, and low-maintenance closed-loop solution for engineering safety monitoring.

[0062] 2. The present invention realizes the accurate monitoring of the displacement of the vertical retaining structure by obtaining the tilt angle data of the retaining structure, can ensure accurate measurement under complex working conditions, and at the same time preprocesses the tilt angle data by using digital filtering technology, which can effectively remove the noise and interference in the data and further improve the measurement accuracy. Combined with the dynamic Kalman filtering algorithm, it can accurately calculate the current motion attitude angle of the module in a complex dynamic environment to achieve high-precision monitoring.

[0063] 3. The present invention uploads the displacement value and motion attitude angle of the vertical retaining structure to the remote data platform, and transmits the monitoring data to the remote server in real time and stably. Engineering management personnel can access the monitoring data at any time and place through terminal devices such as computers or mobile phones without having to visit the site. Moreover, the remote data platform also has an intelligent risk prediction and early warning function. Once the monitoring data exceeds the preset safety threshold, it will immediately notify relevant personnel in multiple ways such as text messages and emails so that timely measures can be taken. And the remote data platform allows management personnel to remotely adjust the parameter settings of the device to achieve refined management of the monitoring process and further improve the intelligent level of project management. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0065] Figure 1 It is a schematic block diagram of an automatic monitoring device for the displacement of a vertical retaining structure according to an embodiment of the present invention.

[0066] Figure 2 It is one of the specific implementation schematic diagrams of an automatic monitoring device for the displacement of a vertical retaining structure according to an embodiment of the present invention;

[0067] Figure 3 It is the second specific implementation schematic diagram of an automatic monitoring device for the displacement of a vertical retaining structure according to an embodiment of the present invention;

[0068] Figure 4 It is the flowchart of an automatic monitoring method for the displacement of a vertical retaining structure according to an embodiment of the present invention.

[0069] In the figure:

[0070] 1. Data acquisition unit; 2. Displacement conversion unit; 3. Monitoring and warning unit; 4. Battery management unit. Specific implementation manner

[0071] To further illustrate the embodiments, the present invention provides drawings, which are part of the disclosure of the present invention. They are mainly used to illustrate the embodiments and can be used to explain the operating principle of the embodiments in conjunction with the relevant descriptions in the specification. With reference to these contents, those of ordinary skill in the art should be able to understand other possible implementation manners and the advantages of the present invention.

[0072] According to an embodiment of the present invention, an automatic monitoring device for the displacement of a vertical retaining structure is provided.

[0073] Now, the present invention will be further described in conjunction with the drawings and specific implementation manners. As Figure 1 shown, the automatic monitoring device for the displacement of a vertical retaining structure according to an embodiment of the present invention includes: a data acquisition unit 1, a displacement conversion unit 2, a monitoring and warning unit 3, and a battery management unit 4.

[0074] The data acquisition unit 1 is used to configure monitoring nodes, obtain sensor data monitored at the top of the retaining wall, and preprocess the sensor data to obtain effective inclination data.

[0075] In the description of the present invention, the data acquisition unit 1 includes: a data acquisition module, a gyroscope preprocessing module, an acceleration preprocessing module, a geomagnetic data preprocessing module, and a data integration module.

[0076] The data acquisition module is used to obtain sensor data monitored by various types of sensors in real time; the sensor data includes gyroscope data, accelerometer data, and geomagnetic sensor data.

[0077] Specifically, high-precision gyroscopes, accelerometers, and geomagnetic sensors are used to real-time sense the tilt angle data of the retaining wall top, namely gyroscope data, accelerometer data, and geomagnetic sensor data. Among them, the gyroscope data is used to measure the dynamic angular velocity; the accelerometer data calculates the static tilt angle through the gravity component; the geomagnetic sensor data is used for direction angle calibration. The measurement accuracy is 0.1 degrees in a dynamic environment and 0.05 degrees in a static environment.

[0078] For example, when heavy rain causes the soil to loosen and the top of the wall begins to slowly tilt to the right, at this time, the gyroscope senses "turning right at 0.01 degrees per second" immediately, and at the same time, the accelerometer detects the deviation of the gravity direction and confirms "static tilt of 0.5 degrees". The geomagnetic sensor finds that "the overall direction is 0.5 degrees east of the original record", and the geomagnetic field sensor locks the absolute direction to detect the absolute orientation of the retaining wall.

[0079] The gyroscope preprocessing module is used to eliminate the static offset error of the gyroscope data based on an adaptive high-pass digital filter to obtain effective gyroscope data.

[0080] Specifically, the static offset error of the gyroscope data is eliminated through an adaptive high-pass digital filter, and the effective signal of the dynamic angular velocity is retained, thereby effectively smoothing random noise, reducing the volatility of the angular velocity data, improving the measurement stability, and ensuring that the tilt angle measurement noise is suppressed within 0.1 degrees in a dynamic environment.

[0081] The accelerometer preprocessing module is used to perform time-domain smoothing processing on the accelerometer data based on a moving average filter, and use a second-order Butterworth low-pass filter to separate the smoothed accelerometer data to obtain effective accelerometer data.

[0082] Specifically, a moving average filter is used for the accelerometer data, and its average value is calculated as the acceleration value at the current moment to perform time-domain smoothing processing on the original data, filter out short-term high-frequency interference, and further combine a second-order Butterworth low-pass filter with a cut-off frequency of 5Hz to effectively separate the gravity component and vibration noise, ensuring that the calculation accuracy of the static tilt angle reaches 0.05 degrees.

[0083] The geomagnetic data preprocessing module is used to eliminate noise from the geomagnetic sensor data based on a windowed median filter, identify the environmental magnetic field interference frequency band through fast Fourier transform frequency-domain analysis, and combine a notch filter to suppress the noise in the interference frequency band to obtain effective geomagnetic sensor data.

[0084] Specifically, a windowed median filter is used for the geomagnetic sensor data to eliminate pulse noise, identify the environmental magnetic field interference frequency band through fast Fourier transform frequency-domain analysis, use a notch filter to suppress the noise in a specific frequency band, and retain the low-frequency geomagnetic signal required for direction angle calibration.

[0085] In the description of the present invention, identifying the environmental magnetic field interference frequency band through fast Fourier transform frequency domain analysis includes:

[0086] Step S101: Normalize the geomagnetic sensor data after noise cancellation, and perform frequency domain transformation on the normalized geomagnetic sensor data using the fast Fourier transform function to obtain the geomagnetic sensor data represented in the frequency domain.

[0087] Step S102: According to the geomagnetic sensor data represented in the frequency domain, obtain the frequency corresponding to each complex number point, and calculate the amplitude spectrum of each frequency point.

[0088] Step S103: Based on the amplitude spectrum of each frequency point, draw an amplitude spectrum diagram, and obtain the frequency corresponding to the peak point with a higher amplitude as the environmental magnetic field interference frequency band.

[0089] It should be noted that after removing the impulse noise through a windowed median filter, the geomagnetic sensor data after noise cancellation is normalized, and the data is mapped to a specific range from -1 to 1 for subsequent analysis and processing; the fast Fourier transform (FFT) function is called to perform frequency domain transformation on the preprocessed data to obtain a complex form of frequency domain representation. Each complex number point corresponds to a specific frequency, and then the amplitude spectrum of each frequency point is calculated, that is, the modulus value of the FFT result is taken to determine the intensity of different frequency components. The amplitude spectrum reflects the amplitude size of the signal at each frequency. By drawing an amplitude spectrum diagram, the horizontal axis represents the frequency, and the vertical axis represents the amplitude corresponding to the frequency. The frequency corresponding to the peak point with a higher amplitude is the interference frequency band.

[0090] A data integration module for integrating the processed effective gyroscope data, effective accelerometer data, and effective geomagnetic sensor data as the effective tilt angle data of the retaining wall.

[0091] A displacement conversion unit 2 for using the dynamic Kalman filtering algorithm to identify and optimize the attitude angle, and calculating the displacement value of the vertical retaining structure in combination with the vertical distance from the bottom to the top of the retaining wall.

[0092] In the description of the present invention, the displacement conversion unit 2 includes: an error accumulation processing module, an optimized angle correction module, and a displacement conversion module.

[0093] The error accumulation processing module is used to obtain the optimized tilt angle at the previous moment, and calculate the optimized tilt angle at the current moment in combination with the effective gyroscope data at the current moment.

[0094] In the description of the present invention, the tilt angle data is preprocessed through data processing and filtering to obtain the angle measured by the gyroscope. The gyroscope is good at capturing dynamic changes, dominating the dynamic response and having a high short-term prediction accuracy. However, long-term drift errors will accumulate. Therefore, it is necessary to calculate the optimized tilt angle at the current moment. The calculation formula for the optimized tilt angle at the current moment is as follows:

[0095]

[0096] In the formula, represents the optimized tilt angle at the k-th moment; represents the optimized tilt angle at the previous moment; ω k represents the angular velocity of the gyroscope corresponding to the valid gyroscope data; Δt represents the time interval.

[0097] The optimized angle correction module is used to correct the valid accelerometer data by using the Kalman gain, and through dynamic weight reduction calculation in combination with the optimized tilt calculation at the current moment, the optimized attitude angle is obtained.

[0098] Specifically, the optimized angle correction module is built with an attitude resolver and combines the dynamic Kalman filtering algorithm. It can accurately calculate the current motion attitude angle of the module in a complex dynamic environment and convert the attitude angle into the displacement value of the top of the retaining wall.

[0099] The dynamic Kalman filtering algorithm dynamically adjusts the fusion weights of multi-source data by real-time estimating the noise covariance matrices of the gyroscope, accelerometer, and geomagnetic sensor, and obtains the optimized attitude angle.

[0100] Among them, the noise covariance matrix includes the process noise covariance matrix Q, which is used to describe the uncertainty of the gyroscope measurement, and the observation noise covariance matrix R, which is used to describe the uncertainty of the accelerometer and geomagnetic sensor measurements. The process noise covariance matrix Q and the observation noise covariance matrix R for describing the uncertainty of the geomagnetic sensor measurement are both fixed values. The observation noise covariance matrix R for describing the uncertainty of the accelerometer measurement is adjusted in real time according to the environmental vibration intensity. The formula is as follows:

[0101]

[0102] In the formula, R a represents the observation noise covariance matrix of the accelerometer data; R base represents the static reference noise fixed value; γ represents the sensitivity fixed value; α rms represents the root mean square value of the acceleration vibration;

[0103] The noise covariance matrix determines the weight of the measurement data in the state update through the Kalman gain K. The formula is as follows:

[0104]

[0105] In the formula, H represents the observation matrix H =

[10] ; R represents the observation noise covariance matrix; Q represents the process noise covariance matrix.

[0106] Under high vibration environment, α rms > 0.2g, which makes the Kalman gain K k decrease, so as to reduce the measurement weight of the accelerometer, increase the measurement weight of the gyroscope, and optimize the attitude angle.

[0107] In the description of the present invention, during the data correction stage, the measurement of the accelerometer is corrected in combination with the Kalman gain. The accelerometer and the geomagnetic sensor provide a static reference and a direction reference, which will be interfered in the dynamic environment. Therefore, it is necessary to dynamically reduce the weight affected by vibration noise to obtain the optimized attitude angle. The calculation formula for the optimized attitude angle is:

[0108]

[0109] In the formula, represents the optimized attitude angle; Z accel，k represents the tilt angle of the accelerometer at time k; K k represents the Kalman gain at time k.

[0110] The displacement conversion module is used to obtain the vertical distance from the bottom to the top of the retaining wall measured by the laser rangefinder, and calculate the displacement value of the vertical retaining structure through the conversion of the optimized attitude angle.

[0111] In the description of the present invention, the calculation formula for the displacement value of the vertical retaining structure is:

[0112] ΔL = h·tanθ k ;

[0113] In the formula, ΔL represents the displacement value of the vertical retaining structure; h represents the vertical distance from the bottom to the top of the retaining wall.

[0114] In the description of the present invention, the monitoring and warning unit 3 includes: an edge communication and computing unit, a digital modeling and display unit, a spatio-temporal correlation prediction unit, and a warning decision support unit.

[0115] The edge communication and computing unit is used to converge the edge computing data to the remote data platform by adopting a wireless communication protocol, so as to realize the data aggregation of distributed monitoring nodes.

[0116] Specifically, this unit realizes the two-way data transmission between the monitoring node and the remote platform by integrating the dual-mode communication protocol of the low-power wide area network (LoRa) and the cellular network (4G / 5G).

[0117] In outdoor scenarios without stable network coverage, the LoRa protocol is preferentially adopted for long-distance and low-rate data upload. When abnormal displacement is detected or a warning is triggered, it automatically switches to the high-speed cellular network to transmit critical data. The edge computing terminal is built with a lightweight data compression algorithm (such as Huffman coding), which boosts the original data compression rate to over 70%, and ensures transmission security through an encryption chip (supporting national cryptography algorithms). When the network is disconnected, the local storage module can cache data for at least 30 days, and resume transmission from the breakpoint through the breakpoint resumption mechanism after the network is restored, ensuring continuous monitoring.

[0118] For example, in the monitoring of mountain retaining walls, this unit can real-time aggregate the displacement and inclination data of multiple nodes, compress them and upload them to the cloud, significantly reducing communication energy consumption and latency.

[0119] The digital modeling display unit is used to construct a three-dimensional virtual model of the retaining wall. By inputting the vertical retaining structure displacement values of each monitoring node, it can real-time map the deformation state of the physical entity.

[0120] Specifically, based on the Building Information Model (BIM) and Geographic Information System (GIS) fusion engine, this unit maps the displacement values of each monitoring node to the corresponding positions in the three-dimensional virtual model, generating a deformation heat map with millimeter-level accuracy.

[0121] The model supports dynamic incremental updates, only reconstructing the locally deformed meshes, reducing computational resource consumption. By combining with drone oblique photography data, it can overlay and display the comparison between the actual terrain and the virtual deformation trend.

[0122] For example, when the vertical displacement of a certain section of the retaining wall exceeds 5mm, the model automatically marks this area in red and generates a historical displacement curve graph. In addition, it supports Augmented Reality (AR) interaction. Engineers can view the deformation data superimposed on the real scene by scanning the site with a mobile terminal, assisting in quickly locating risk points.

[0123] The spatio-temporal correlation prediction unit is used to predict the future deformation trend of the retaining wall and identify potential risk points based on historical and real-time data through spatio-temporal information fusion and mining.

[0124] It should be noted that spatio-temporal fusion consists of spatial fusion and temporal fusion. Spatial fusion is based on the Graph Convolutional Network (GCN), which adaptively aggregates features such as displacement and strain of adjacent monitoring points, capturing the local collaborative effects in the overall deformation of the retaining wall; temporal fusion uses the Bidirectional Gated Recurrent Unit (Bi-GRU) to model the long-term dependencies of the displacement sequence along the time axis, and combines the attention mechanism to achieve the dynamic coupling of spatio-temporal features.

[0125] Assume that the monitoring data has time-varying Gaussian noise. The model outputs the mean and variance of displacement prediction, replacing traditional point prediction. The loss function of maximum likelihood estimation is used to quantify prediction uncertainty, enhancing the robustness to noise. With the help of the full-process training mechanism, during model training, the hidden layer states at each time step are involved in loss calculation, avoiding information loss caused by traditional truncated training and strengthening the model's overall learning ability for the deformation accumulation process.

[0126] In the displacement monitoring of vertical retaining structures, through spatial correlation modeling, GCN is used to dynamically assign weights to the features of multiple measurement points, solving the problem that traditional single-measurement-point models ignore spatial correlation. For example, when a section of retaining wall undergoes chain deformation due to local settlement, GCN automatically enhances the feature transmission of relevant nodes through the weights of the adjacency matrix, accurately capturing the deformation propagation path.

[0127] For time series dependence and long-term prediction, through Bi-GRU and full-process training, historical displacement data and external environmental parameters (such as rainfall, groundwater level) are fused to solve the problem that traditional models have inaccurate predictions of long-term deformation trends. For example, the soil softening effect caused by continuous heavy rain can be predicted 72 hours in advance through time series dependence. In addition, for improving noise robustness, the probability prediction framework directly models the sensor noise in the monitoring data (such as fiber optic wavelength drift, visual target recognition error), and quantifies the prediction credibility through variance output, avoiding false alarms caused by noise interference.

[0128] In the description of the present invention, based on historical and real-time data, through spatio-temporal information fusion and mining, predicting the future deformation trend of the retaining wall and identifying potential risk points includes:

[0129] Step S31: Obtain the historical and real-time vertical retaining structure displacement values, sensor data, and environmental data (such as rainfall, groundwater level, temperature, etc.) of the retaining wall, map all monitoring nodes to a unified spatio-temporal coordinate, and use the Pearson correlation coefficient to quantify the spatial correlation between each monitoring node, screen strongly correlated neighbor nodes, and construct an adjacency matrix.

[0130] Specifically, the vertical displacement, strain, and inclination data are collected in real time through a sensor network deployed on the retaining wall, and the environmental parameters of the weather station (rainfall) and hydrological monitoring equipment (groundwater level, temperature) are synchronously accessed. Historical data is extracted from the database, covering at least one complete hydrological year cycle (12 months) of monitoring records.

[0131] For missing value processing, linear interpolation is used to fill short-term missing values (such as short-term sensor failures), and long-term missing values are marked as invalid. For noise filtering, a moving average filter (window width Δt = 1 hour) is applied to the displacement data, and finally, data with different dimensions such as displacement and rainfall are normalized to the [0, 1] interval.

[0132] All monitoring nodes are installed with Beidou RTK positioning modules (accuracy ±1 cm) to obtain longitude and latitude coordinates (EPSG: 4326) and convert them into a local engineering coordinate system (such as UTM projection); the NTP protocol is used to synchronize the clocks of each node to ensure that the data timestamp error < 1 ms and eliminate the timing misalignment caused by asynchronous sampling.

[0133] For each pair of measurement points (i, j), calculate the Pearson correlation coefficient of their historical displacement sequences and set a threshold ρ thres = 0.8. If ρ ij ≥ ρ thres , it is determined as a strongly correlated neighbor. For each measurement point, retain the first k = 5 neighbors with the highest correlation coefficients.

[0134] Step S32: Based on the adjacency matrix, aggregate the features of adjacent monitoring nodes through a multi-layer graph convolutional network, normalize the adjacency matrix, and layer by layer transfer the node features to capture the synergistic effect of the local deformation of the retaining wall.

[0135] Specifically, aggregating the features of adjacent monitoring nodes through a multi-layer graph convolutional network and normalizing the adjacency matrix can solve the problem of feature transfer deviation caused by the difference in node degrees in traditional graph convolution and avoid gradient explosion or disappearance. It is necessary to add an identity matrix IN to the diagonal of the original adjacency matrix A, so that each node retains its own features during feature aggregation (for example, the displacement value of a certain measurement point will not be completely covered by the influence of neighbors), and then use the degree matrix D (the diagonal elements are the degrees of the nodes, that is, the number of neighbors) to normalize the adjacency matrix. If measurement point i has 3 neighbors and measurement point j has 5 neighbors, the feature contribution weights of the two will be balanced after normalization, avoiding nodes with more neighbors from dominating the entire network.

[0136] By stacking multiple GCN layers, gradually expand the feature receptive field to capture the local-to-global synergistic deformation. The operations of a single GCN layer include the following aspects:

[0137] Input: Node feature matrix (N is the number of measurement points, F is the feature dimension);

[0138] Output: Updated feature matrix H (l+1) , and the formula is: H (l+1) = σ(AH (l) W (l) );

[0139] Among them, represents the trainable weight matrix of the l-th layer.

[0140] In addition, for the multi-layer stacking example: The first layer: Aggregate the direct neighbor features to capture the local deformation near the measurement point (such as the synchronous change in the displacement of adjacent points due to cracks in a certain section of the wall). The second layer: Indirectly aggregate through the neighbors of the neighbors to perceive the deformation correlation in a larger range (such as the upstream displacement causing downstream chain settlement).

[0141] With the coupling effect of the adjacency matrix weight and feature transfer, the node features are transferred layer by layer to capture the synergy effect of the local deformation of the retaining wall, which can be used to show the local uplift deformation in a certain area of the retaining wall due to the sudden increase in soil pressure. The implementation process includes:

[0142] 1. Adjacency matrix weight: The weight of strongly correlated neighbors (ρ ij > 0.8) is relatively high. For example, the weight a AB of measurement points A and B is 0.9 because they are physically close and have the same displacement trend.

[0143] 2. Feature aggregation: The features (displacement, strain) of measurement point A are transferred to the feature expression of measurement point B through high weights, enhancing the correlation of their co-deformation.

[0144] 3. Multi-layer transfer: The second-layer GCN further transfers the features of measurement point B to a farther measurement point C, thereby spreading the influence of local deformation to the entire monitoring network.

[0145] Step S33: Input the spatially aggregated features into a bidirectional gated recurrent network to model the long-term dependence of the displacement sequence along the time axis, and introduce a spatio-temporal cross-attention mechanism to calculate the correlation weight between the spatial features and the temporal hidden state, generating a coupled state vector that fuses spatio-temporal dynamics.

[0146] Specifically, the bidirectional gated recurrent network (Bi-GRU) models the temporal dependence, which can capture the long-term forward and backward dependence relationships of the displacement sequence and solve the gradient vanishing problem of the traditional RNN.

[0147] By dynamically correlating the spatial features and the temporal hidden state, the feature expression of key spatio-temporal nodes is enhanced. Its calculation steps include the following aspects:

[0148] 1. Feature mapping: Map the spatial features and the temporal hidden state to the query, key, and value spaces respectively.

[0149] 2. Attention score: Calculate the correlation weight of each measurement point i at time t. The calculation formula is:

[0150]

[0151] In the formula, α ti represents the correlation weight of each monitoring node i at time t; d k represents the attention dimension; Qi represents the query vector of the monitoring node; K t represents the key vector at time t.

[0152] 3. Feature weighted fusion: Generate a spatio-temporal coupling feature vector, and the calculation formula is:

[0153]

[0154] In the formula, c t represents the spatio-temporal coupling feature vector; V i represents the value vector of monitoring node i.

[0155] Example: When the displacement of a certain measuring point surges due to the rising groundwater level, its associated weight α ti increases, enhancing the influence of this node on the overall prediction.

[0156] Step S34: Based on the spatio-temporal fusion features, output the mean difference and variance of the displacement prediction value through a linear layer, construct a time-varying Gaussian analysis model, optimize the model parameters using the maximum likelihood loss function, and use the time-varying Gaussian analysis model to predict the mean and variance of the deformation amount at future moments.

[0157] In the description of the present invention, the parameter prediction formula of the time-varying Gaussian analysis model is:

[0158]

[0159] In the formula, μ t represents the mean value of the predicted displacement value of the vertical retaining structure; W μ represents the linear layer weight of the mean value; c t represents the spatio-temporal fusion feature vector; b μ represents the mean bias term; σ t represents the variance of the predicted displacement value of the vertical retaining structure; W σ represents the linear layer weight of the variance; b σ represents the variance bias term.

[0160] The formula of the maximum likelihood loss function is:

[0161]

[0162] In the formula, L represents the maximum likelihood loss function; y t represents the actual observed value of the displacement of the vertical retaining structure at time t; T represents the total length of the time series.

[0163] Step S35: According to the prediction results of the mean value and variance, evaluate the deformation trend and potential risk points of the retaining wall, and introduce an online learning mechanism to optimize the model using the latest monitoring data.

[0164] In the description of the present invention, according to the mean and variance prediction results, the deformation trend and potential risk points of the retaining wall are evaluated, and an online learning mechanism is introduced to optimize the model using the latest monitoring data, including:

[0165] Step S351: The future deformation amount of the retaining wall follows a Gaussian distribution. According to the mean and variance prediction results, the probability interval in which the future deformation amount is located is judged, and the risk points are calibrated based on the deformation amounts of each monitoring node.

[0166] Specifically, the future deformation amount follows a Gaussian distribution The model output is the probability distribution parameter rather than a single value. For example: when the model predicts that the μ of a certain measuring point t = 5.2 mm, and σ t = 0.8 mm, its 95% confidence interval is 5.2 ± 1.96 × 0.8 mm, that is, [3.6 mm, 6.8 mm]. If the upper limit of this interval exceeds the preset safety threshold (such as 6 mm), it is determined as a potential risk point. The risk level is dynamically divided according to the proximity of the mean to the threshold:

[0167] Yellow warning: μ t Between 50% and 80% of the threshold (such as 3 mm to 4.8 mm);

[0168] Orange warning: μ t Between 80% and 100% of the threshold (such as 4.8 mm to 6 mm);

[0169] Red warning: μ t Exceeds the threshold and σ t < 1 mm.

[0170] The calibration of risk points needs to consider the spatial correlation. If more than 3 adjacent measuring points trigger warnings simultaneously, it is determined as a regional systematic risk.

[0171] Step S352: Calculate the mutual information value between the environmental factors and the displacement value of the vertical retaining structure, dynamically adjust the input weights of the time-varying Gaussian distribution model, and construct an environment-threshold mapping table to match the current environmental adjustment standard.

[0172] Specifically, the correlation between the environmental factors (such as rainfall R t , groundwater level W t ) and the displacement is quantified by weighted mutual information:

[0173]

[0174] Wherein, the weight w(x, y) is defined according to spatio-temporal proximity (for example, the weight of recent data is higher). p(x, y) represents the joint probability distribution of X and Y (such as the joint distribution of displacement and environmental factors); p(x) and p(y) represent the marginal probability distributions of X and Y (separate probability distributions).

[0175] Construct an "environment-threshold" mapping table. For example: when R t > 50 mm, the displacement threshold is lowered from 6 mm to 4.5 mm; when W t exceeds the warning water level, the warning sensitivity of the associated measuring points is increased by 20%.

[0176] Step S353: Set an update period to regularly collect the latest monitoring data, calculate the loss value of the new data through the maximum likelihood loss function. If the loss value is greater than the preset threshold, update the parameters of the time-varying Gaussian distribution model.

[0177] Step S354: Map the mean prediction result to a three-dimensional virtual model to visually display the risk distribution of the retaining wall.

[0178] An early warning decision support unit is used to set a dynamic multi-level early warning strategy, trigger multi-level early warnings and generate emergency disposal and maintenance suggestions according to the analysis and prediction results.

[0179] Specifically, based on a dynamic three-level early warning mechanism, this unit matches the prediction results with real-time data to preset thresholds. For example, the thresholds can be set as follows: a yellow warning (displacement rate > 2 mm / day) triggers a platform pop-up window and SMS reminder; an orange warning (cumulative displacement > 10 mm and abnormal earth pressure) activates an audible and visual alarm and generates preliminary reinforcement suggestions (such as adding drainage ditches); a red warning (topological instability risk) then links to the emergency management system to push the anchor reinforcement plan and evacuation route.

[0180] At the same time, the pre-plan matching engine screens the disposal records of similar working conditions from the historical case library (such as the cable anchor reinforcement plan in the treatment of a landslide in a certain place in Fujian), combines blockchain technology to record the early warning events and disposal processes, and realizes responsibility traceability.

[0181] For example, when the predicted risk probability of a flood wall overturning exceeds 85%, automatically call the emergency material allocation plan of adjacent projects and synchronize the disposal instructions to the mobile terminals of on-site personnel.

[0182] In the present invention, the spatio-temporal fusion model can achieve millimeter-level displacement monitoring accuracy, output a confidence interval in combination with probability prediction, and support three-level early warnings of yellow / orange / red. For example, when the displacement mean value of a certain area exceeds the threshold and the variance is low, a red early warning is directly triggered and a reinforcement plan is pushed. It has long-term stability and adaptability. The solar power supply and intelligent power consumption management unit ensure the continuous operation of the device in a gridless environment; the online learning mechanism regularly fine-tunes the model parameters with the latest data to adapt to long-term evolutions such as the aging of retaining wall materials and the change of earth pressure.

[0183] It supports full automation and low maintenance. The sensor zero point is calibrated daily through the self-check system, and automatically switches to the redundant channel in case of anomalies; the maintenance alarm unit generates maintenance work orders in combination with GPS positioning, reducing the frequency of manual inspections and significantly reducing the operation and maintenance costs in complex scenarios such as remote mountainous areas and flood control walls along rivers.

[0184] In the description of the present invention, the battery management unit 4 includes: an energy supply module, a status monitoring module, a power consumption regulation module, and a maintenance alarm module.

[0185] The energy supply module is used to provide continuous power for the monitoring device by combining solar power generation and battery energy storage, and adjust the energy input and output in real time to prevent overcharging or over-discharging from damaging the battery.

[0186] The status monitoring module is used to continuously track the real-time power, health status, and working temperature of the battery, collect data through voltage and current sensors, and analyze the remaining battery life and performance degradation trend.

[0187] The power consumption regulation module is used to dynamically adjust the device power consumption according to the current power and monitoring requirements. In the low-power state, non-core functions are automatically turned off, and only the basic sensors are retained for operation.

[0188] The maintenance alarm module is used to establish a multi-level early warning mechanism, and grade and prompt maintenance requirements through methods such as platform pop-up windows and SMS notifications, automatically generate maintenance work orders and push them to designated personnel.

[0189] Please refer to Figure 4 , the present invention also provides an automatic monitoring method for the displacement of a vertical retaining structure, and the monitoring method includes:

[0190] S1. Configure distributed monitoring nodes on the top of the retaining wall, obtain the sensor data monitored on the top of the retaining wall, and preprocess the sensor data to obtain effective inclination data;

[0191] S2. Based on the effective inclination data, use the dynamic Kalman filtering algorithm to identify and optimize the attitude angle, and combine the vertical distance from the bottom to the top of the retaining wall to calculate the displacement value of the vertical retaining structure;

[0192] S3. Upload the displacement values of the vertical retaining structure calculated by all monitoring nodes to the remote data platform, construct a monitoring network, analyze and predict the deformation trend of the retaining wall, and trigger an alarm according to the analysis results.

[0193] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments.

[0194] As Figures 2 to 3 shown, this embodiment provides an automatic monitoring device for the displacement of a vertical retaining structure, which is installed on the top of the retaining wall. In this technical solution, the sensor uses high-precision gyroscopes, accelerometers, and geomagnetic sensors to continuously sense the tilt angle of the top of the retaining wall;

[0195] Among them, the gyroscope data is used to measure the dynamic angular velocity; the accelerometer data calculates the static tilt angle through the gravity component; the geomagnetic sensor data is used for direction angle calibration; the measurement accuracy is 0.1 degrees in a dynamic environment and 0.05 degrees in a static environment. And the sensor uses MEMS sensors, which have high-precision monitoring, environmental adaptability, long battery life, and the ability to resist vibration interference.

[0196] Specifically, the gyroscope can use the BMI088 gyroscope, which has a high dynamic response ability and is suitable for capturing instantaneous vibrations or deformations of the retaining wall; it has an industrial temperature range (-40°C to 85°C) and can adapt to extreme outdoor environments. The triaxial accelerometer can use the ADI ADXL355 accelerometer, which has ultra-low noise characteristics, ensures that the calculation accuracy of the static tilt angle reaches 0.05°, has an internal anti-aliasing filter to suppress high-frequency vibration interference, and supports solar power generation to extend the battery life. The geomagnetic sensor can use the MMC5983MA geomagnetic sensor, which has high resolution and supports sub-degree heading angle calibration (0.1° accuracy); it has an anti-interference design and is suitable for the retaining wall structure;

[0197] As Figures 2 to 3 shown, in this technical solution, data preprocessing and filtering are performed on the tilt angle data measured by the sensor to reduce measurement noise and improve measurement accuracy, which is applicable to various environments.

[0198] Among them, for the gyroscope data, the static offset error is removed through an adaptive high-pass digital filter, and the effective signal of the dynamic angular velocity is retained, thereby effectively smoothing the random noise, reducing the volatility of the angular velocity data, improving the measurement stability, and ensuring that the tilt measurement noise is suppressed within 0.1 degrees in a dynamic environment. For the accelerometer data, a moving average filter is used to calculate its average value as the acceleration value at the current moment, and the original data is smoothed in the time domain to filter out short-term high-frequency interference. Further combined with a second-order Butterworth low-pass filter with a cut-off frequency of 5 Hz, the gravity component and vibration noise are effectively separated, ensuring that the calculation accuracy of the static tilt angle reaches 0.05 degrees. A windowed median filter is used to eliminate impulse noise, and the environmental magnetic field interference frequency band is identified through fast Fourier transform frequency domain analysis. A notch filter is used to suppress the noise in the specific frequency band, and the low-frequency geomagnetic signal required for direction angle calibration is retained.

[0199] In this technical solution, the monitored data after preprocessing and filtering preprocessing is obtained through an attitude resolver, and the current optimized attitude angle (motion attitude angle) of the module is calculated by combining the dynamic Kalman filtering algorithm. Through specific formulas, the attitude angle is converted into the displacement value of the top of the retaining wall using a programming script.

[0200] Among them, the attitude resolver is connected to a high-precision microelectromechanical gyroscope, a three-axis accelerometer, and a geomagnetic sensor through a dedicated digital bus to achieve multi-channel synchronous data acquisition and transmission, and uses the dynamic Kalman filtering algorithm to optimize the attitude angle. Automatically improve the credibility of the gyroscope in a vibrating environment, and preferentially use the gravity component of the accelerometer for calculation in a static scenario, ensuring that the measurement accuracy is 0.1 degrees in a dynamic environment and 0.05 degrees in a static environment.

[0201] The dynamic Kalman filtering algorithm dynamically adjusts the fusion weight of multi-source data by real-time estimating the noise covariance matrix of the gyroscope, accelerometer, and geomagnetic sensor, and obtains the optimized attitude angle to accurately calculate the current motion attitude angle of the module in a complex dynamic environment.

[0202] As an optional implementation, a programming script or data processing software can be used to complete the calculation of trigonometric functions through specific formulas for the optimized current optimized attitude angle (motion attitude angle) to calculate the displacement value of the top of the retaining wall.

[0203] In addition, this technical solution has communication and remote management functions. It is equipped with a 4G communication device, which can transmit the real-time monitored data to the remote data platform in a safe and stable manner, realizing cloud data management and analysis for monitoring the displacement value of the top of the retaining wall. The communication device also has the function of receiving instructions, and can remotely calibrate the sensor and adjust the device setting parameters through the built-in microcontroller to ensure the accuracy and reliability of the monitored data.

[0204] The remote data platform supports real-time alert functions. When abnormal displacement or tilt angle exceeds the safe range, it can send alerts to users or maintenance personnel in the first place. The remote management function inside the device requires a built-in microcontroller to parse the instruction content, identify the types of sensors to be operated, and remotely modify the sensor sampling frequency, sensitivity, or alarm threshold to adapt to different working conditions.

[0205] In addition, in this technical solution, the battery management module is configured with an ultra-long endurance battery that supports solar power generation to be applicable to long-term monitoring in outdoor environments, ensuring the long-term stable operation of the monitoring device. It also has a battery status monitoring function to regularly report the battery power and ensure timely reminders to maintenance personnel for maintenance or battery replacement in case of low battery. The battery management module supports solar charging, optimizes the efficiency of the photovoltaic panel through an MPPT controller, and adopts a battery management system (BMS) to prevent overcharging, over-discharging, short circuits, and abnormal temperatures, and regularly sends the battery status to the cloud platform through a 4G module.

[0206] As an alternative implementation, the ultra-long endurance battery can use lithium thionyl chloride batteries, which have extremely high energy density, are suitable for extreme temperatures, and have extremely low self-discharge rates, making them suitable for long-term outdoor deployments over 10 years and are widely used in Internet of Things nodes and remote monitoring devices.

[0207] In summary, by means of the above technical solutions of the present invention, the present invention can accurately sense and monitor the displacement of the top of the retaining wall through the tilt angle data, solve the defect that traditional monitoring means cannot obtain displacement data in real time and accurately, adopt an attitude resolver and a dynamic Kalman filtering algorithm, and perform remote data transmission and control through a remote data platform, so as to realize automatic monitoring, and the monitoring data is more comprehensive and accurate, avoiding complex manual measurement; at the same time, edge computing is combined to perform real-time noise reduction and spatio-temporal alignment on the data, significantly improving the data reliability; based on the spatio-temporal prediction model of the graph convolutional network and the recurrent neural network, the overall deformation trend and local risk points of the retaining wall can be dynamically analyzed, the uncertainty can be quantified through probability prediction and a hierarchical early warning can be triggered; it can support long-term continuous operation in an outdoor environment without power grid, and combined with a self-check mechanism and a remote maintenance interface, the frequency of manual inspections can be greatly reduced; the modular design takes into account both the deployment convenience and the system scalability, providing a high-precision, fully automatic, and low-maintenance closed-loop solution for engineering safety monitoring. The present invention realizes the precise monitoring of the displacement of the vertical retaining structure by obtaining the tilt angle data of the retaining structure, can ensure accurate measurement under complex working conditions, and at the same time preprocesses the tilt angle data by using digital filtering technology, which can effectively remove the noise and interference in the data and further improve the measurement accuracy; combined with the dynamic Kalman filtering algorithm, it can accurately calculate the current motion attitude angle of the module in a complex dynamic environment to achieve high-precision monitoring. The present invention uploads the displacement value and motion attitude angle of the vertical retaining structure to the remote data platform, and transmits the monitoring data to the remote server in real time and stably; engineering management personnel can access the monitoring data at any time through terminal devices such as computers or mobile phones without having to be on site; moreover, the remote data platform also has an intelligent risk prediction and early warning function. Once the monitoring data exceeds the preset safety threshold, it will immediately notify relevant personnel in various ways such as text messages and emails so that timely measures can be taken; and the remote data platform allows management personnel to remotely adjust the parameter settings of the device to achieve refined management of the monitoring process and further improve the intelligent level of project management.

[0208] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An automatic monitoring device for the displacement of a vertical retaining structure, characterized in that, Including: A data acquisition unit, configured to configure monitoring nodes, obtain sensor data monitored at the top of the retaining wall, and preprocess the sensor data to obtain effective inclination data; A displacement conversion unit, configured to use a dynamic Kalman filtering algorithm to identify and optimize the attitude angle, and calculate the displacement value of the vertical retaining structure in combination with the vertical distance from the bottom to the top of the retaining wall; A monitoring and warning unit, configured to upload the displacement value of the vertical retaining structure to a remote data platform, construct a monitoring network, analyze and predict the deformation trend of the retaining wall, and trigger a warning according to the analysis result; A battery management unit, configured to provide battery state monitoring and management, and regularly report the battery power.

2. The automatic displacement monitoring device for a vertical retaining structure according to claim 1, characterized in that, The data acquisition unit includes: A data acquisition module, configured to obtain sensor data monitored in real time by various types of sensors; the sensor data includes gyroscope data, accelerometer data, and geomagnetic sensor data; A gyroscope preprocessing module, configured to eliminate the static offset error of the gyroscope data based on an adaptive high-pass digital filter to obtain effective gyroscope data; An accelerometer preprocessing module, configured to perform time-domain smoothing on the accelerometer data based on a moving average filter, and perform separation processing on the smoothed accelerometer data using a second-order Butterworth low-pass filter to obtain effective accelerometer data; A geomagnetic data preprocessing module, configured to eliminate noise from the geomagnetic sensor data based on a windowed median filter, identify the environmental magnetic field interference frequency band through fast Fourier transform frequency-domain analysis, and suppress the noise in the interference frequency band in combination with a notch filter to obtain effective geomagnetic sensor data; A data integration module, configured to integrate the processed effective gyroscope data, effective accelerometer data, and effective geomagnetic sensor data as the effective inclination angle data of the retaining wall.

3. The automatic displacement monitoring device for a vertical retaining structure according to claim 2, characterized in that, The identification of the environmental magnetic field interference frequency band through fast Fourier transform frequency-domain analysis includes: Performing normalization processing on the geomagnetic sensor data after noise elimination, and performing frequency-domain transformation on the normalized geomagnetic sensor data using a fast Fourier transform function to obtain the geomagnetic sensor data represented in the frequency domain; According to the geomagnetic sensor data represented in the frequency domain, obtaining the frequency corresponding to each complex number point, and calculating the amplitude spectrum of each frequency point; Based on the amplitude spectrum of each frequency point, plotting an amplitude spectrum diagram, and obtaining the frequency corresponding to the peak point with a higher amplitude as the environmental magnetic field interference frequency band.

4. The automatic monitoring device for the displacement of a vertical retaining structure according to claim 2, characterized in that, The displacement conversion unit includes: An error accumulation processing module, configured to obtain the optimized inclination angle at the previous moment, and calculate the optimized inclination angle at the current moment in combination with the effective gyroscope data at the current moment; An optimized angle correction module, configured to correct the effective accelerometer data using the Kalman gain, and calculate the optimized attitude angle through dynamic weight reduction calculation in combination with the optimized inclination calculation at the current moment; A displacement conversion module, configured to obtain the vertical distance from the bottom to the top of the retaining wall measured by a laser rangefinder, and calculate the displacement value of the vertical retaining structure through conversion of the optimized attitude angle.

5. The automatic monitoring device for the displacement of a vertical retaining structure according to claim 4, characterized in that, The calculation formula for the optimized inclination angle at the current moment is: In the formula, represents the tilt optimization angle at time k; represents the tilt optimization angle at the previous time; ω k represents the gyroscope angular velocity corresponding to the valid gyroscope data; Δt represents the time interval; The calculation formula for the optimized attitude angle is: In the formula, represents the optimized attitude angle; Z accel，k represents the accelerometer tilt angle at time k; K k represents the Kalman gain at time k; The calculation formula for the displacement value of the vertical retaining structure is: ΔL = h·tanθ k ; In the formula, ΔL represents the displacement value of the vertical retaining structure; h represents the vertical distance from the bottom to the top of the retaining wall.

6. The automatic displacement monitoring device for a vertical retaining structure according to claim 1, characterized in that, The monitoring and early warning unit includes: An edge communication and computing unit, which is used to converge edge computing data to a remote data platform by using a wireless communication protocol to achieve data aggregation of distributed monitoring nodes; A digital modeling and display unit, which is used to construct a three-dimensional virtual model of the retaining wall and, by inputting the displacement values of the vertical retaining objects at each monitoring node, map the deformation state of the physical entity in real time; A spatio-temporal correlation prediction unit, which is used to predict the future deformation trend of the retaining wall and identify potential risk points based on historical and real-time data through spatio-temporal information fusion and mining; An early warning decision support unit, which is used to set a dynamic multi-level early warning strategy, trigger multi-level early warnings according to the analysis and prediction results, and generate emergency disposal and maintenance suggestions.

7. An automatic monitoring device for the displacement of a vertical retaining structure according to claim 6, characterized in that, The predicting the future deformation trend of the retaining wall and identifying potential risk points based on historical and real-time data through spatio-temporal information fusion and mining includes: Obtaining the historical and real-time displacement values of the vertical retaining structure, sensor data and environmental data of the retaining wall, mapping all monitoring nodes to a unified spatio-temporal coordinate, and using the Pearson correlation coefficient to quantify the spatial correlation between each monitoring node, screening strongly correlated neighbor nodes, and constructing an adjacency matrix; Based on the adjacency matrix, aggregating the features of adjacent monitoring nodes through a multi-layer graph convolutional network, normalizing the adjacency matrix, and transmitting node features layer by layer to capture the synergistic effect of local deformation of the retaining wall; Inputting the spatially aggregated features into a bidirectional gated recurrent network, modeling the long-term dependence relationship of the displacement sequence along the time axis, introducing a spatio-temporal cross-attention mechanism, calculating the correlation weight between the spatial features and the temporal hidden state, and generating a coupled state vector that fuses spatio-temporal dynamics; Based on the spatio-temporal fusion features, outputting the mean difference and variance of the displacement prediction value through a linear layer, constructing a time-varying Gaussian analysis model, optimizing the model parameters by using the maximum likelihood loss function, and predicting the mean and variance of the deformation amount at a future moment by using the time-varying Gaussian analysis model; According to the mean and variance prediction results, evaluating the deformation trend and potential risk points of the retaining wall, and introducing an online learning mechanism to optimize the model by using the latest monitoring data.

8. An automatic displacement monitoring device for a vertical retaining structure according to claim 7, characterized in that, The parameter prediction formula of the time-varying Gaussian analysis model is: where μ t represents the mean value of the predicted displacement value of the vertical retaining structure; W μ represents the linear layer weight of the mean value; c t represents the spatio-temporal fusion feature vector; b μ represents the mean bias term; σ t represents the variance of the predicted displacement value of the vertical retaining structure; W σ represents the linear layer weight of the variance; b σ represents the variance bias term; The formula of the maximum likelihood loss function is: In the formula, L represents the maximum likelihood loss function; y t represents the actual observed value of the displacement of the vertical retaining structure at time t; T represents the total length of the time series.

9. The automatic displacement monitoring device for a vertical retaining structure according to claim 7, characterized in that, The evaluating the deformation trend and potential risk points of the retaining wall according to the mean and variance prediction results, and introducing an online learning mechanism to optimize the model by using the latest monitoring data includes: The future deformation amount of the retaining wall follows a Gaussian distribution. According to the mean and variance prediction results, judge the probability interval where the future deformation amount is located, and calibrate the risk points according to the deformation amount of each monitoring node; Calculating the mutual information value between the environmental factors and the displacement value of the vertical retaining structure, dynamically adjusting the input weight of the time-varying Gaussian distribution model, and constructing an environment-threshold mapping table to match the current environment adjustment standard; Setting an update period to regularly collect the latest monitoring data, calculating the loss value of the new data through the maximum likelihood loss function, and if the loss value is greater than the preset threshold, updating the parameters of the time-varying Gaussian distribution model; Mapping the mean prediction result to the three-dimensional virtual model to visually display the risk distribution of the retaining wall.

10. An automatic monitoring method for the displacement of a vertical retaining structure, which is used to implement the automatic monitoring device for the displacement of the vertical retaining structure described in any one of claims 1-9, characterized in that, The monitoring method includes: Distributed monitoring nodes are configured at the top of the retaining wall to obtain the sensor data of the top of the retaining wall, and the sensor data is preprocessed to obtain effective inclination data; Based on the effective inclination data, the dynamic Kalman filtering algorithm is used to identify and optimize the attitude angle, and combined with the vertical distance from the bottom to the top of the retaining wall, the displacement value of the vertical retaining structure is calculated; The displacement values of the vertical retaining structures calculated by all monitoring nodes are uploaded to the remote data platform to construct a monitoring network, analyze and predict the deformation trend of the retaining wall, and trigger an early warning according to the analysis results.

Images