Anchorage diaphragm wall measuring point anomaly identification and global deformation three-dimensional visualization method and system

Abstract

The invention discloses an anchorage diaphragm wall measuring point abnormity identification and global deformation three-dimensional visualization method and system, and the method comprises the steps: arranging displacement sensors on an anchorage diaphragm wall, forming a uniformly-distributed three-dimensional monitoring array, and building an anchorage diaphragm wall monitoring system comprising the three-dimensional monitoring array; the network time of the anchorage diaphragm wall monitoring system is synchronized; collecting data of various sensors on the underground diaphragm wall in real time and uploading the data to a cloud monitoring database; dividing the diaphragm wall into a grid structure, expanding discrete monitoring data into high-density grid data through an interpolation algorithm, and realizing global high-precision displacement simulation of the diaphragm wall; abnormal data can be automatically eliminated; and the deformation of the diaphragm wall is three-dimensionally visualized. According to the method, the overall deformation state of the anchorage diaphragm wall is reflected more accurately, abnormal data are eliminated through an abnormal value detection mode, and the reliability of monitoring data is improved; the problem that the space deformation characteristics of the diaphragm wall are difficult to directly reflect by traditional monitoring data is solved.

Description

Technical Field

[0001] The present invention relates to the field of civil engineering monitoring and visualization technology, and in particular to a method, system, and storage medium for identifying abnormalities at measuring points of anchored diaphragm walls and for three-dimensional visualization of global deformation. Background Art

[0002] Deep foundation pit projects are a crucial component of bridge construction. With the continuous advancement of urbanization, the scale and number of deep foundation pit projects are increasing. Underground diaphragm walls are the most widely used form of support structure for deep foundation pits, offering advantages such as convenient construction, effective water-stopping, and high rigidity. However, ground-connected walls are also one of the most risky components of deep foundation pit projects. Their stress state is complex, and their deformation directly impacts the safety of the entire foundation pit. Traditional automated monitoring methods for ground-connected walls primarily use inclinometer tube integration to determine horizontal displacement. This method can capture internal deformation within the wall, but the discrete distribution of monitoring points makes it difficult to fully reflect the overall deformation state of the ground-connected wall. Furthermore, the monitoring data lacks intuitive visualization, making it difficult for engineers to quickly assess the structural safety.

[0003] In recent years, with the rapid development of new-generation information technologies such as the Internet of Things, big data, and cloud computing, the use of inclinometers to collect structural deformation data in real time and then analyze data to achieve all-round, dynamic monitoring of ground-connected walls has become a research hotspot in the field of deep foundation pit engineering. However, traditional flat display panels of monitoring data are difficult to intuitively reflect the spatial deformation characteristics of ground-connected walls, and data anomalies often affect decision-making during the construction process. There is an urgent need to develop a three-dimensional visualization method suitable for ground-connected wall structures that can transform massive amounts of monitoring data into intuitive and vivid three-dimensional scenes, providing engineering personnel with more intelligent and efficient safety management methods. This study aims to realize key technologies such as multi-source heterogeneous data fusion, real-time performance monitoring, and three-dimensional visual presentation of ground-connected walls, providing new solutions for the safety of deep foundation pit engineering. Summary of the Invention

[0004] In response to the above defects or improvement needs of the existing technology, the present invention aims to solve the problem that traditional monitoring data plane display panels are difficult to intuitively reflect the spatial deformation characteristics of the ground-connected wall and are difficult to comprehensively reflect the overall deformation state of the ground-connected wall. At the same time, the monitoring data lacks intuitive visual display, and the construction process often has abnormal data status that affects decision-making.

[0005] The present invention provides a method for identifying anomalies at measuring points of an anchored diaphragm wall and visualizing the global deformation in three dimensions, comprising the following steps:

[0006] Step S1. Sensor Arrangement: Displacement sensors are arranged on the anchor diaphragm wall to form a uniformly distributed three-dimensional monitoring array in the horizontal and vertical directions, and an anchor diaphragm wall monitoring system including the three-dimensional monitoring array is established;

[0007] Step S2. Network time synchronization of the anchorage-ground-wall monitoring system: Time synchronization of each node of the anchorage-ground-wall monitoring system in step S1;

[0008] Step S3. Data collection and upload: Real-time data collection from various sensors on the ground and wall is uploaded to the cloud monitoring database;

[0009] Step S4. Global displacement simulation of the ground-connected wall: (Global means the entire ground-connected wall): Thin plate spline (TPS) method is used to fit the discrete monitoring data by minimizing the energy functional to obtain the observed value r(x i ,y i ,z i ) corresponding estimation function f; dividing the ground-connected wall into a grid structure, and expanding the discrete monitoring data into high-density grid data of the entire domain through the estimation function f, thereby realizing virtual sensing of the ground-connected wall surface;

[0010] Step S5. Abnormal sensor data removal: Perform residual analysis on the high-density grid data in step S4 to identify and remove abnormal sensor data. If there is an abnormality, remove the abnormal data. If there is no abnormality, proceed directly to the next step.

[0011] Step S6. 3D visualization of ground-connected wall deformation: Render the data processed in step S5 into an interactive 3D heat map with color overlays for various operations and information queries.

[0012] Furthermore, the specific method of step S2 includes: establishing a unified time base, synchronizing the system clocks of each displacement sensor node, aggregation node, and database server through the NTP (Network Time Protocol) protocol, and realizing time synchronization of the distributed monitoring network.

[0013] Furthermore, the specific method of step S3 includes: collecting data from various sensors on the ground-connected wall in real time, using wireless sensor network technology based on the ZigBee protocol to encapsulate the monitoring data into a unified data frame format, and uploading it to the cloud monitoring database in real time.

[0014] Furthermore, step S3 also includes dynamically adjusting the sampling frequency of each sensor according to the data change rate and signal characteristics, thereby reducing data redundancy and energy consumption while ensuring information integrity.

[0015] Furthermore, step S3 also includes adding a timestamp tag to the data frame to record the time of data collection. According to the sampling frequency differences of different sensors, an adaptive data collection strategy is designed.

[0016] Furthermore, in step S4, the method of fitting discrete monitoring data by minimizing the energy functional using the Thin Plate Spline (TPS) method includes:

[0017] (1) At a certain time point, suppose the ground-connected wall has n sensor observation values ​​r(x i ,y i ,z i ):

[0018] r(x i ,y i ,z i )=f(x i ,y i ,z i )+ε(x i ,y i ,z i ),i=1,2,...,n (1)

[0019] Where f is the observed value r(x i ,y i ,z i ) corresponds to the estimation function, ε is the estimation error, which is a zero-mean, non-correlated random error; (x i ,y i ,z i ) is the Euclidean space coordinate;

[0020] (2) To make the estimated function f as close to the observed value as possible, the energy function ξ(f) should be as small as possible:

[0021] ξ(f)=E d (f)+λE(f) (2)

[0022]

[0023] Where λ is a smoothing parameter that can balance the control deformation stiffness E d (f) and the smoothness E(f) of the fitted surface.

[0024] Furthermore, the method of dividing the ground-connected wall into a grid structure in step S4 and expanding the discrete monitoring data into high-density grid data by estimating the function f includes: dividing the ground-connected wall into a grid structure in a virtual coordinate system, using the estimation function f The displacement value of each grid point is calculated, and the grid data is uploaded to the cloud database of the anchored ground-connected wall monitoring system for archiving in a timely manner, realizing high-density virtual sensing on the surface of the ground-connected wall.

[0025] Furthermore, the method for eliminating abnormal sensor data in step S5 includes:

[0026] (1) Estimate the values ​​of all monitoring points based on the thin plate spline interpolation function f established in step S3;

[0027] (2) Compare the estimated value with the measured value point by point and calculate the residual error of each monitoring point;

[0028] (3) mean μ and standard deviation σ of the distribution statistics residual;

[0029] (4) The monitoring points whose absolute value of residual exceeds μ + 2σ are marked as suspected abnormal points;

[0030] (5) Eliminate the identified abnormal sensor data, record the abnormal sensors, and reconstruct the thin plate spline interpolation function f;

[0031] (6) Repeat the process (1)-(5) until no new abnormal points appear, and finally confirm the interpolation function and the abnormal sensor list.

[0032] Furthermore, the three-dimensional visualization method of the ground-connected wall deformation in step S6 includes:

[0033] (1) Return the result of obtaining data to the front end in JSON format;

[0034] (2) The front end sends a request through AJAX or Fetch.API to obtain the (x, y, z) coordinates of the monitoring point and its corresponding deformation value;

[0035] (3) Use Plotly.js to render a 3D heat map;

[0036] (4) On the heat map, a red mark is further superimposed at each monitoring point to indicate the location of the sensor;

[0037] (5) Use the scatter3d layer to render sensor points separately, clearly mark the sensor positions, and support mouse interaction;

[0038] (6) The content displayed during hovering is customized through hoverinfo and hovertemplate, so that detailed information of each sensor can be queried.

[0039] As another aspect of the present invention, it also relates to an anchored ground-connected wall abnormality identification and three-dimensional visualization device, comprising an anchored ground-connected wall monitoring system network time synchronization unit, a data acquisition and upload unit, a discrete monitoring data expansion unit, an abnormal sensor data elimination unit, and a ground-connected wall deformation three-dimensional visualization unit;

[0040] The network time synchronization unit of the anchored ground-connected wall monitoring system is used to synchronize the time of each node of the anchored ground-connected wall monitoring system;

[0041] The data acquisition and uploading unit is used to collect, process and store the raw sensor data of the anchored ground-connected wall monitoring system;

[0042] The discrete monitoring data expansion unit is used to fit the discrete monitoring data by minimizing the energy functional using the Thin Plate Spline (TPS) method to obtain the observation value r(x i ,y i ,z i ) corresponding to the estimation function f; the ground-connected wall is divided into a grid structure, and the discrete monitoring data is expanded into high-density grid data through the estimation function f to realize virtual sensing of the ground-connected wall surface;

[0043] The abnormal sensor data elimination unit is used to perform residual analysis on the high-density grid data of the anchorage diaphragm wall monitoring system, and identify and eliminate abnormal sensor data;

[0044] The three-dimensional visualization unit for deformation of the anchored ground-connected wall is used to construct a three-dimensional visualization model of the anchored ground-connected wall monitoring system.

[0045] As another aspect of the present invention, it also relates to a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to provide a method for monitoring and three-dimensionally visualizing deformation of diaphragm walls in large foundation pit projects.

[0046] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0047] (1) The method for identifying abnormalities at measuring points and visualizing the global deformation of anchored diaphragm walls of the present invention is a three-dimensional detection array method for constructing anchored diaphragm wall structures. Compared with the traditional method of obtaining horizontal displacement by mainly using inclinometer tube integration in automated monitoring of anchored diaphragm walls, the method provides a comprehensive monitoring method that can more accurately reflect the overall deformation state of the anchored diaphragm wall.

[0048] (2) The method for identifying abnormalities at measuring points of anchored diaphragm walls and visualizing global deformation in three dimensions of the present invention optimizes the data collection process and reduces energy consumption by dynamically adjusting the sampling frequency during the collection and processing of massive data collected from the structure of anchored diaphragm walls. It also eliminates abnormal data through two outlier detection methods, thereby improving the reliability of monitoring data and avoiding the problem of data abnormalities affecting decision-making during the construction process.

[0049] (3) The present invention's method for identifying anomalies at measurement points and visualizing global deformation of anchored diaphragm walls, by constructing a method for visualizing the deformation of anchored diaphragm walls, can transform massive amounts of monitoring data into intuitive 3D scenes, providing engineers with a more intelligent and efficient safety management tool. This solves the problem that traditional flat display panels of monitoring data are unable to intuitively reflect the spatial deformation characteristics of diaphragm walls. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a schematic diagram of the overall process of a preferred embodiment of the present invention;

[0051] Figure 2 Schematic top view of a three-dimensional monitoring array for anchored diaphragm walls according to a preferred embodiment of the present invention;

[0052] Figure 3 This is a schematic diagram of the inclinometer casing side of a preferred embodiment of the present invention;

[0053] Figure 4 This is a schematic diagram showing the comparison of whether or not there are abnormal points in the data during abnormal point detection in a preferred embodiment of the present invention;

[0054] Figure 5 A schematic diagram comparing the time-dependent changes in the three-dimensional behavior of the horizontal displacement structure of the ground-connected wall in a preferred embodiment of the present invention; DETAILED DESCRIPTION

[0055] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only intended to illustrate the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0056] Please refer to Figure 1-Figure 3 The embodiment of the present invention provides a method for identifying abnormalities at measuring points of anchored diaphragm walls and visualizing their global deformation in three dimensions, including:

[0057] (1) Multi-source heterogeneous sensor deployment:

[0058] Various types of displacement sensors are rationally arranged on the ground-connected wall (fiber grating sensors, for example, can also be used to improve monitoring accuracy and stability, while enabling distributed monitoring and obtaining more comprehensive deformation information), forming a uniformly distributed three-dimensional monitoring array in the horizontal and vertical directions. In this embodiment, taking a circular foundation pit as an example, the ground-connected wall of the circular foundation pit is evenly spaced with no fewer than 12-20 holes, for example, 16 holes, namely P1 to P16. Inclinometer tubes are placed in each hole. The depth of the inclinometer tubes is consistent with the ground-connected wall structure. The guide groove of the inclinometer tube is perpendicular to the edge of the foundation pit and points to the center of the foundation pit. The sensor acquisition frequency is once per hour.

[0059] (2) Data collection, uploading, and spatiotemporal data interpolation and expansion:

[0060] Establish a unified time base and synchronize the system clocks of each sensor node with the aggregation node and database server through the NTP (Network Time Protocol) protocol to achieve time synchronization of the distributed network; in some other embodiments, high-precision satellite timing technology can also be used to ensure that the time error of each sensor node is controlled within a smaller range.

[0061] Real-time data collection from various sensors on the ground-connected wall is performed. Using wireless sensor network technology based on the ZigBee protocol, the monitoring data is encapsulated into a unified data frame format and uploaded to a cloud-based monitoring database in real time. In some embodiments, an automatic backup module can be configured during the data collection and upload process. This module monitors the data transmission status in real time and immediately activates local cache storage upon detecting a network failure or other situation that may result in data loss. The collected data is stored in a local reliable storage device in a specific format and timestamp, while also recording the relevant data collection node information and original path information. Once the network returns to normal, the locally cached data is re-uploaded to the cloud-based monitoring database in chronological order and using a data integrity verification mechanism. Data comparison and fusion are then performed to ensure data continuity and integrity.

[0062] A timestamp is added to the data frame to record the moment of data acquisition. An adaptive data acquisition strategy is designed to address the differences in sampling frequencies among different sensors. Based on the data change rate and signal characteristics, the sampling frequency of each sensor is dynamically adjusted to ensure information integrity while reducing data redundancy and energy consumption.

[0063] Please refer to Figure 4 The raw data uploaded to the monitoring database may contain outliers and noise, requiring preliminary processing to improve data quality. Outlier detection is performed using the statistically based 3σ principle. For each monitoring indicator, the mean μ and standard deviation σ of the time series data are calculated. Data points outside the range of μ ± 3σ are marked as outliers and replaced by linear interpolation of normal values ​​from the preceding and following moments. The 3σ principle, based on the assumption that data follow a normal distribution, can identify the vast majority of random anomalies.

[0064] For a small number of deterministic anomalies (such as continuous anomalies caused by sensor failure), the next step is to leverage correlations and constraints between monitored variables to further improve anomaly identification accuracy. The Thin Plate Spline (TPS) method is introduced to fit discrete monitoring data by minimizing an energy functional to generate virtual sensor data. Compared to other interpolation methods, TPS can better reflect the physical characteristics of surface changes and is smooth and continuous, making it suitable for discrete point interpolation.

[0065] Assume that at a certain time point, the ground-connected wall has n sensor observation values ​​r(x i ,y i ,z i ), which can be expressed as follows:

[0066] r(x i ,y i ,z i )=f(x i ,y i ,z i )+ε(x i ,y i ,z i ),i=1,2,...,n (1)

[0067] Where f is the observed value r(x i ,y i ,z i ) corresponds to the estimation function, ε is the estimation error, which is a zero-mean, non-correlated random error; (x i ,y i ,z i ) are Euclidean coordinates.

[0068] If we want to make the estimated function f as close to the observed value as possible, the energy function ξ(f) should be as small as possible:

[0069] ξ(f)=E d (f)+λE(f) (2)

[0070]

[0071] Where λ is a smoothing parameter that can balance the control deformation stiffness E d (f) and the smoothness E(f) of the fitted surface.

[0072] Please refer to Figure 5In this embodiment, the ground-connected wall can be divided into multiple sections vertically and circumferentially, for example, 25 vertical sections and 180 circumferential sections. The grid size is approximately 2.5m x 2.5m. Using the estimation function f, the displacement values ​​of 24*179=4296 grid points can be calculated, achieving high-density virtual sensing of the ground-connected wall surface. After the grid data calculation is completed, it is promptly uploaded to the database for archiving. Figure 5 After the discrete monitoring data is expanded into global high-density grid data (left side is before implementation, right side is after implementation), the color distribution of displacement changes spreads to the entire ground-connected wall.

[0073] (3) Abnormal sensor data removal

[0074] The values ​​of all monitoring points are estimated by estimating the function f, and the estimated values ​​are compared with the measured values ​​point by point to calculate the residual ε(x i ,y i ,z i ).

[0075] The mean μ and standard deviation σ of the statistical residuals are distributed, and the monitoring points with the absolute value of the residual exceeding μ+2σ are marked as suspected outliers. The data of the identified abnormal sensors are discarded, and the estimation function f (thin plate spline interpolation function) is reconstructed. The residual distribution of all monitoring points is recalculated until no new outliers appear. Finally, the interpolation function and the list of abnormal sensors are determined.

[0076] In some other embodiments, multi-dimensional statistical analysis can also be performed on the collected sensor data. First, statistical quantities such as the mean, median, and mode of each data item are calculated to obtain the data's central tendency characteristics. Second, the variance and standard deviation of the data are analyzed to assess the degree of data dispersion. Then, for data from different time periods, a data distribution histogram or frequency distribution curve is constructed to observe the data distribution pattern. By setting a reasonable threshold range, based on statistical quantities and distribution characteristics, data points outside the normal range are identified. Combined with methods such as residual analysis, anomalies in the sensor data can be comprehensively judged to improve the accuracy and reliability of anomaly identification.

[0077] (4) 3D interactive operation

[0078] To achieve 3D web visualization of ground-anchored wall deformation, we retrieved grid point prediction data and measured data from monitoring points from the database and returned the query results to the front-end in JSON format. The front-end then sent a request via AJAX or the Fetch API to obtain the (x, y, z) coordinates of the monitoring points and their corresponding deformation values. Then, we used Plotly.js to render a 3D heat map.

[0079] Specifically, the 3D model is drawn using the mesh3d layer in Plotly.js. The intensity property maps the deformation value of each monitoring point to a color, and the rendering is done using a continuous color spectrum (Viridis). Users can rotate, zoom, and pan the 3D model to view different angles. The system also provides interactive features, allowing users to hover or click a point to display the specific displacement data.

[0080] Based on the heat map, the system can further overlay red markers on the location of each monitoring point to indicate the location of the sensor. By using the scatter3d layer to render these sensor points separately, the red dots not only clearly mark the sensor locations but also support mouse interaction. When the user moves the mouse over a red dot, a pop-up display will show the sensor number and corresponding monitoring point information. To enhance the interactive experience, the system can customize the content displayed during hovering through hoverinfo and hovertemplate, allowing users to quickly query detailed information for each sensor.

[0081] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for identifying abnormal points of anchored diaphragm walls and three-dimensional visualization of global deformation, characterized by: The steps include: Step S1. Arrange displacement sensors on the anchored ground-connected wall to form a three-dimensional monitoring array evenly distributed in the horizontal and vertical directions, and establish an anchored ground-connected wall monitoring system including the three-dimensional monitoring array; Step S2. Time synchronization of each node of the anchorage ground-connected wall monitoring system in step S1; Step S3. Collect data from various sensors on the ground and walls in real time and upload them to the cloud monitoring database; Step S4. Use the thin plate spline interpolation method to fit the discrete monitoring data by minimizing the energy functional to obtain the observed value r(x i ,y i ,z i ) corresponding to the estimation function f; dividing the ground-connected wall into a grid structure, expanding the discrete monitoring data into high-density grid data through the estimation function f, and realizing virtual sensing of the ground-connected wall surface; Step S5. Perform residual analysis on the high-density grid data in step S4 to identify and eliminate abnormal sensor data; Step S6. Render the data processed in step S5 into an interactive three-dimensional heat map with superimposed color markings to perform various operations and information queries.

2. The method for identifying abnormalities at measuring points of anchored diaphragm walls and visualizing their global deformation in three dimensions according to claim 1 is characterized by: The specific method of step S2 includes: Establish a unified time base, synchronize the system clocks of each displacement sensor node with the aggregation node and database server through the NTP protocol, and realize the time synchronization of the distributed monitoring network.

3. The method for anomaly identification and global deformation three-dimensional visualization of anchored diaphragm wall measuring points according to claim 1 is characterized by: The specific method of step S3 includes: The data of various sensors on the ground-connected wall are collected in real time, and the monitoring data is encapsulated into a unified data frame format using wireless sensor network technology based on the ZigBee protocol, and uploaded to the cloud monitoring database in real time.

4. The method for identifying abnormalities at measuring points of anchored diaphragm walls and visualizing deformation in three dimensions according to claim 3 is characterized by: Step S3 also includes adding a timestamp tag to the data frame to record the data collection time, and designing an adaptive data collection strategy according to the sampling frequency differences of different sensors.

5. The method for identifying abnormalities at measuring points of anchored diaphragm walls and visualizing their global deformation in three dimensions according to any one of claims 1 to 4, characterized in that: In step S4, the thin plate spline interpolation method is used to fit the discrete monitoring data by minimizing the energy functional, including: S41: At a certain time point, suppose the ground-connected wall has n sensor observation values ​​r(x i ,y i ,z i ): r(x i ,y i ,z i )=f(x i ,y i ,z i )+ε(x i ,y i ,z i ),i=1,2,…,n (1) Where f is the observed value r(x i ,y i ,z i ) corresponds to the estimation function; ε is the estimation error, which is a zero-mean, non-correlated random error; (x i ,y i ,z i ) is the Euclidean space coordinate; S42: To make the estimated function f as close to the observed value as possible, the energy function ξ(f) should be: ξ(f)=E d (f)+λE(f) (2) Where λ is a smoothing parameter that can balance the deformation stiffness E d (f) and the smoothness E(f) of the fitted surface.

6. The method for identifying abnormal points of anchored diaphragm walls and visualizing their global deformation in three dimensions according to claim 5, characterized in that: In step S4, the ground-connected wall is divided into a grid structure, and the method of expanding the discrete monitoring data into high-density grid data by estimating the function f includes: dividing the ground-connected wall into a grid structure in a virtual coordinate system, calculating the displacement value of each grid point by using the estimation function f, and uploading the grid data to the cloud database of the anchored ground-connected wall monitoring system for archiving in a timely manner after the calculation is completed; realizing high-density virtual sensing of the surface of the ground-connected wall.

7. The method for identifying abnormalities at measuring points of anchored diaphragm walls and visualizing their global deformation in three dimensions according to any one of claims 1 to 4, characterized in that: The method for eliminating abnormal sensor data in step S5 includes: (1) According to the estimation function established in step S3 f , estimate the values ​​of all monitoring points; (2) Compare the estimated value with the measured value point by point and calculate the residual error of each monitoring point; (3) mean μ and standard deviation σ of the distribution statistics residual; (4) The monitoring points whose absolute values ​​of residuals exceed μ+2σ are marked as suspected abnormal points; (5) Eliminate the identified abnormal sensor data, record the abnormal sensors, and reconstruct the estimation function f; (6) Repeat the process (1)-(5) until no new abnormal points appear, and finally confirm the interpolation function and the list of abnormal sensors.

8. The method for anomaly identification and global deformation three-dimensional visualization of anchored diaphragm wall measuring points according to any one of claims 1 to 4, characterized in that: The three-dimensional visualization method of ground-connected wall deformation in step S6 includes: (1) Return the result of obtaining data to the front end in JSON format; (2) The front end sends a request through AJAX or Fetch.API to obtain the (x, y, z) coordinates of the monitoring point and its corresponding deformation value; (3) Use Plotly.js to render a 3D heat map; (4) On the heat map, a red mark is further superimposed at each monitoring point to indicate the location of the sensor; (5) Use the scatter3d layer to render sensor points separately, clearly mark the sensor locations, and support mouse interaction; (6) The content displayed during hovering can be customized through hoverinfo and hovertemplate, so that detailed information of each sensor can be queried.

9. An anchored ground-connected wall anomaly recognition and three-dimensional visualization system, characterized by: Network time synchronization unit, used to synchronize the time of each node of the anchorage ground-connected wall monitoring system; Data acquisition and upload unit, used to collect, process and store the raw sensor data of the anchored ground-connected wall monitoring system; The discrete monitoring data expansion unit is used to fit the discrete monitoring data by minimizing the energy functional using the thin plate spline interpolation method to obtain the observed value r(x i ,y i ,z i ) corresponding to the estimation function f; dividing the ground-connected wall into a grid structure, expanding the discrete monitoring data into high-density grid data through the estimation function f, and realizing virtual sensing of the ground-connected wall surface; The abnormal sensor data elimination unit is used to perform residual analysis on the high-density grid data of the anchored ground-connected wall monitoring system, and identify and eliminate abnormal sensor data; The three-dimensional visualization unit of the ground-anchored wall deformation is used to construct a three-dimensional visualization model of the anchored ground-anchored wall monitoring system.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the method for identifying abnormalities at measuring points of anchored diaphragm walls and three-dimensional visualization of global deformation as described in any one of claims 1 to 8.

Images