Sluice structure safety monitoring system and method

Abstract

The invention discloses a water gate structure safety monitoring system and method, belongs to the technical field of computer communication, breaks through isolated parameter monitoring limitation, constructs multi-physics field dynamic coupling analysis capability, carries out deep fusion processing on multi-source data, improves data reliability and defect identification precision in a complex environment, and improves the safety of the water gate structure. And the use effect of real-time quantitative evaluation of the structural state is realized by utilizing the technical means of dynamic model inversion and parameter calibration, that is, a stress-deformation-leakage three-dimensional finite element coupling model is constructed, dynamic calibration of material parameters and real-time inversion of the structural state are realized in combination with long-term monitoring data, the limitation of isolation of traditional monitoring is broken through, and the real-time quantitative evaluation of the structural state is realized. And a more accurate technical support is provided for water gate safety assessment.

Description

Technical Field

[0001] The present invention relates to the field of computer communication technology, and more particularly to a sluice structure safety monitoring system and method. Background Art

[0002] As a key control hub in water conservancy projects, the structural safety of sluice gates is directly related to the stable operation of important functions such as flood control, irrigation, and navigation. Traditional sluice gate safety monitoring systems mainly rely on independent data collection from stress sensors, deformation monitoring devices, and leakage detection equipment, and achieve safety warnings through threshold comparison. However, existing technologies have significant drawbacks:

[0003] On the one hand, monitoring parameters are isolated. Traditional methods monitor key indicators such as stress, deformation, and leakage as independent parameters, ignoring the dynamic coupling relationship between the three during the operation of the sluice. For example, stress concentration in the sluice body may cause local deformation, which in turn causes damage to the anti-seepage structure and induces leakage. The development of leakage will further aggravate the abnormal stress distribution, forming a complex evolution process of multi-physical field interaction. However, existing technologies fail to establish a linkage analysis model between the three, making it difficult to accurately reveal the inherent evolution mechanism of structural safety, resulting in one-sided monitoring results and low early warning reliability.

[0004] On the other hand, the depth of multi-source data fusion is insufficient. Although existing technologies have achieved the synchronous collection of stress, deformation, leakage and other data, as well as optical detection of surface defects (such as ordinary image shooting), data processing still remains at the level of single parameter filtering, threshold judgment or simple image recognition, and has not built a multi-physics field coupling model that includes mechanical response, deformation characteristics and seepage characteristics. For example, traditional deformation monitoring (such as total stations and laser displacement meters) only provides geometric displacement data, and does not form a correlation analysis with stress field distribution and seepage pressure changes. Leakage detection is also mostly limited to single-point monitoring of flow or pressure, and does not combine the structural stress state to judge the actual threat level of leakage to the safety of the gate body;

[0005] Furthermore, there are gaps in model inversion technology. Existing sluice safety assessments rely on empirical formulas or simplified mechanical models, lacking the ability to calibrate dynamic parameters and conduct multi-field coupling inversion based on measured data. For example, the stress-deformation relationship is only linearly fitted using the material elastic modulus, without considering the nonlinear properties of concrete and the effect of seepage on the degradation of material parameters. Furthermore, leakage path analysis is not linked to the stress and strain distribution of the sluice body, making it difficult to quantify the extent of structural damage in real time.

[0006] In response to the above problems, the present invention proposes a sluice structure safety monitoring system and method that integrates multi-parameter dynamic coupling relationships. By constructing a three-dimensional finite element coupling model of stress-deformation-leakage and combining long-term monitoring data, dynamic calibration of material parameters and real-time inversion of structural status are achieved, breaking through the isolated limitations of traditional monitoring and providing more accurate technical support for sluice safety assessment. Summary of the Invention

[0007] 1. Technical problems to be solved

[0008] In response to the problems existing in the prior art, the purpose of the present invention is to provide a sluice structure safety monitoring system and method. It constructs a stress-deformation-leakage three-dimensional finite element coupling model, combines long-term monitoring data to realize dynamic calibration of material parameters and real-time inversion of structural status, breaks through the isolated limitations of traditional monitoring, and provides more accurate technical support for sluice safety assessment.

[0009] 2. Technical solution

[0010] To solve the above problems, the present invention adopts the following technical solutions.

[0011] A sluice structure safety monitoring system includes a data acquisition module, a data transmission module, a data processing module, a model building module, an early warning module and a polarization imaging module;

[0012] The data acquisition module is used to collect stress data, deformation data, leakage data and surface image data of the sluice, and specifically includes:

[0013] Stress sensors, using resistance strain sensors or vibrating wire sensors, are deployed at key stress-bearing locations including but not limited to sluice piers and gate support beams to collect stress and strain signals in real time;

[0014] The deformation monitoring device includes a total station and a laser displacement meter. The total station is set at the connecting section of the two banks to monitor the relative displacement data of the top of the sluice body, the bottom of the pier and the connecting section of the two banks. The laser displacement meter is set at the top of the sluice body, the bottom of the pier and the connecting section of the two banks to monitor the horizontal displacement, vertical settlement and relative displacement data of the structure surface.

[0015] Leakage detection devices, using leakage pressure sensors or flow sensors, are placed downstream of the anti-seepage curtain at the gate foundation, around the drainage holes and gate waterstops to monitor seepage pressure and leakage flow;

[0016] The polarization imaging module includes:

[0017] Polarized light source, using linearly polarized light source or circularly polarized light source, emits incident light of specific polarization state (0°, 45°, 90°, 135° linear polarization) to evenly illuminate the concrete surface of the sluice gate;

[0018] Polarization camera, equipped with a multi-channel polarization filter, capable of simultaneously capturing images in at least two different polarization states, with a pixel resolution of no less than 20 megapixels;

[0019] The data transmission module adopts an industrial-grade Ethernet or 4G / 5G wireless transmission module, is connected to the data acquisition module and the polarization imaging module, and transmits the stress data, deformation data, leakage data and polarization image data to the data processing module in real time through the TCP / IP protocol or the MQTT protocol;

[0020] The data processing module includes a digital signal processing unit and an image analysis unit. The digital signal processing unit uses a Kalman filter algorithm to filter the received data. The Kalman filter algorithm equation is:

[0021] X k|k =X k|k-1 +K k (Z k -H k X k|k-1 )

[0022] Among them, X k|k is the optimal estimated state at time k, X k|k-1 is the predicted state at time k based on time k_1, K k is the Kalman gain, Z k is the measured value at time k, H k For the measurement matrix, the cubic spline interpolation algorithm is used to fill in the missing data, and the Z-score method is used to identify and eliminate abnormal data outside 3σ;

[0023] The image analysis unit sequentially performs median filtering noise reduction and histogram equalization enhancement processing on the polarization image, and uses the Otsu algorithm to automatically determine the threshold T for threshold segmentation. The threshold segmentation formula is:

[0024]

[0025] Among them, g(x,y) is the segmented image, f(x,y) is the original image, and then the defect area is extracted. The characteristic parameters of crack length, width, and direction are obtained through the Canny edge detection algorithm, and the Euclidean distance matching analysis is performed with the preset defect feature library;

[0026] The model building module establishes a three-dimensional finite element model containing the stress-deformation-leakage dynamic coupling relationship based on the sluice structure mechanics model and seepage field theory. The coupling equation of the three-dimensional finite element model is:

[0027]

[0028] Among them, σ is the stress tensor, ρ is the material density, g is the acceleration of gravity, D is the flexibility matrix, ε0 is the initial strain, k is the permeability coefficient, and h is the head function. Then, using historical monitoring data for more than five years, the least squares method is used to calibrate the model material parameters and invert the real-time stress distribution, deformation displacement and leakage path of the sluice.

[0029] The early warning module presets three levels of safety thresholds and compares the model inversion results with the thresholds in real time. When the stress exceeds 80% of the design value, the deformation rate is greater than 2mm / month, or the leakage flow rate is greater than 5L / s, a graded early warning signal is issued through an audible and visual alarm and a remote monitoring platform. At the same time, combined with the results of polarization image analysis, structural defects with crack widths greater than 0.2mm or depths greater than 5cm are located and marked.

[0030] Furthermore, the total station adopts a prism-free ranging mode with a ranging accuracy of ≤±(2mm+2ppm×D), and the laser displacement meter has a measuring range of ≥500mm and a resolution of ≤0.01mm.

[0031] Furthermore, the data transmission module is configured with an industrial-grade fiber optic switch, supports Modbus / TCP and OPC UA industrial protocols, has a data transmission delay of ≤50ms, and a bit error rate of ≤10-9.

[0032] Furthermore, the model building module adopts the ANSYS Mechanical and COMSOL Multiphysics joint simulation platform to achieve parametric modeling of the stress-deformation-seepage coupling field through the APDL language.

[0033] Furthermore, the early warning module integrates the GIS geographic information system to display the location and early warning status of each monitoring point of the sluice in real time, supports remote push of early warning information on the mobile phone APP, and the response time is ≤10s.

[0034] A method for monitoring the safety of a sluice structure comprises the following steps:

[0035] S1. Data acquisition: The stress sensor collects stress and strain signals at a frequency of 10 Hz. The deformation monitoring device automatically measures three times a day at 8:00, 14:00, and 20:00. The leakage sensor monitors the seepage pressure and flow in real time. The polarization imaging module performs a full-coverage scan of the sluice surface every week, with a single scan covering an area ≥ 200 m2 , collecting images of 4 linear polarization states, and the collection angles of the 4 linear polarization state images are 0°, 45°, 90° and 135° respectively;

[0036] S2. Data transmission: The sensor data is packaged into JSON format via industrial Ethernet. The JSON format data includes a timestamp, sensor ID, data value, and checksum. The data is then transmitted to the data processing server via a VPN encrypted channel. The polarization image is uploaded via the FTP protocol. The transmission time for a single image is ≤15s.

[0037] S3. Data processing: The sensor data is sequentially subjected to Kalman filtering, Z-score outlier removal, and cubic spline interpolation. The polarization image is then subjected to 3×3 median filtering for noise reduction and adaptive histogram equalization for enhancement. The Otsu algorithm is used to calculate the optimal threshold T, and the defect area is segmented. Morphological operations are used to remove noise areas with an area of ​​less than 100 pixels, and the crack geometric features are extracted.

[0038] S4. Model inversion: A three-dimensional geometric model is established based on the sluice gate design drawings. A tetrahedral mesh is formed, with a cell size of ≤50 cm. Initial material parameter values ​​are then input, including the concrete elastic modulus, Poisson's ratio, and permeability coefficient. Preprocessed historical data for more than five years is used as a training set. The model parameters are optimized using the Levenberg-Marquardt algorithm. Real-time stress values, deformation, and leakage at each monitoring point of the sluice gate are then inverted.

[0039] S5. Warning judgment: When the inverse stress value ≥ safety threshold, deformation ≥ displacement threshold, or leakage ≥ flow threshold, a warning of the corresponding level is triggered. The safety threshold is the design stress × 0.8, the displacement threshold is 20 mm, and the flow threshold is 5 L / s.

[0040] When the crack width is ≥0.2mm and the length is ≥50cm, it is marked as a structural defect. A maintenance work order is generated based on the location information, and the early warning information and defect report are simultaneously pushed to the sluice management department.

[0041] Furthermore, during model training in step S4, the iteration termination condition is set to root mean square error ≤ 0.5% or the number of iterations ≥ 1000 times, and the parameter calibration accuracy requires elastic modulus error ≤ 5% and permeability coefficient error ≤ 10%.

[0042] Furthermore, in step S1, the scanning path of the polarization imaging module adopts a serpentine coverage strategy, and the overlap rate of adjacent scanning strips is ≥30%, ensuring that surface defects are monitored without blind spots.

[0043] 3. Beneficial effects

[0044] Compared with the prior art, the advantages of the present invention are:

[0045] (1) This solution breaks through the limitations of isolated parameter monitoring and builds a multi-physical field dynamic coupling analysis capability. Traditional technologies only analyze stress, deformation, and leakage data in isolation, while the present invention realizes the linkage inversion of the three for the first time through a three-dimensional finite element coupling model. That is, the model construction module is based on the sluice structure mechanics model and seepage field theory to establish a coupling relationship including stress tensor, deformation displacement, and leakage path. Combined with historical monitoring data for more than 5 years, material parameter calibration is performed, and the interaction between the internal stress distribution, deformation evolution, and seepage path of the structure can be inverted in real time. Compared with the traditional single parameter threshold judgment, this technology can accurately reveal the chain risk evolution process of "stress concentration → structural deformation → leakage induction", avoiding the early warning lag or misjudgment caused by isolated parameter analysis;

[0046] (2) This solution uses deep fusion processing of multi-source data to improve data reliability and defect recognition accuracy in complex environments. The data processing module integrates Kalman filtering algorithm, Z-score outlier elimination and cubic spline interpolation completion technology to solve the problems of data noise interference and missing in traditional monitoring, ensuring the reliability of stress, deformation and leakage data. The polarization imaging module uses four-channel linear polarization light scanning and Otsu threshold segmentation algorithm, combined with Canny edge detection to extract crack geometric features. Compared with traditional visible light imaging, it can identify subtle defects with a width of ≥0.2mm and a depth of ≥5cm. The serpentine scanning strategy is used to achieve blind spot monitoring of surface defects, which improves the convenience of early identification of surface cracks in the concrete of the sluice structure.

[0047] (3) This solution uses dynamic model inversion and parameter calibration to achieve real-time quantitative evaluation of structural status. Different from traditional empirical formulas or simplified models, this invention uses the Levenberg-Marquardt algorithm to optimize the parameters of the three-dimensional finite element model, uses historical data of more than 5 years as a training set, sets iteration termination conditions, and achieves high-precision inversion of the real-time stress value, deformation, and leakage of the sluice. It accurately simulates the evolution process of structural damage under multi-field coupling, and provides a scientific basis for the health status assessment of the sluice. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of the system composition of the present invention;

[0049] Figure 2 This is a schematic diagram of the data acquisition principle of the data acquisition module of the present invention;

[0050] Figure 3 This is a schematic diagram of the data transmission principle of the data transmission module of the present invention;

[0051] Figure 4 This is a schematic diagram of the data processing principle of the data processing module of the present invention;

[0052] Figure 5 This is a schematic diagram of the model construction principle of the model construction module of the present invention;

[0053] Figure 6 This is a schematic diagram of the warning principle of the early warning module of the present invention;

[0054] Figure 7 It is a schematic diagram of the method steps of the present invention;

[0055] Figure 8 This is a schematic diagram of the principle of step S1 of the present invention;

[0056] Figure 9 This is a schematic diagram of the principle of step S2 of the present invention;

[0057] Figure 10 This is a schematic diagram of the principle of step S3 of the present invention;

[0058] Figure 11 This is a schematic diagram of the principle of step S4 of the present invention;

[0059] Figure 12 This is a schematic diagram of the principle of step S5 of the present invention. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the specification of the present invention; it is obvious that the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0061] Example:

[0062] See also Figures 1-12 The present invention constructs a sluice structure safety monitoring system and method, which realizes a full-process technical closed loop from multi-dimensional data acquisition, heterogeneous data transmission, multi-modal processing to multi-physics field coupling inversion and intelligent early warning through the collaborative work of multiple modules. Its core working principle is as follows:

[0063] 1. Principle of multi-dimensional data collection

[0064] 1. Real-time capture of stress data

[0065] Sensor selection and deployment:

[0066] Resistive strain sensors or vibrating wire sensors are embedded in key stress-bearing locations such as sluice piers and gate support beams. Resistive strain sensors are based on the piezoresistive effect of metal strain gauges. When the structure is deformed by stress, the resistance of the strain gauge changes linearly with the strain. This is converted into a voltage signal through a Wheatstone bridge circuit, and dynamic stress and strain signals are collected in real time at a frequency of 10 Hz. These sensors are suitable for monitoring high-frequency stress fluctuations.

[0067] The vibrating wire sensor uses the positive correlation between the vibration frequency of the steel wire and the stress it is subjected to (frequency formula:

[0068] Where T is tension, L is string length, and ρ is line density). It obtains frequency signals through electromagnetic excitation and vibration pickup, has anti-electromagnetic interference capabilities, and is suitable for long-term stable monitoring.

[0069] 2. Double-precision monitoring of deformation data

[0070] Total station macro positioning:

[0071] A prism-free total station (distance measurement accuracy ≤±(2mm+2ppm×D)) is fixedly installed at the connecting section on both sides of the sluice gate. Using the triangulation method, the three-dimensional coordinates of the preset monitoring points on the top of the gate body, the bottom of the gate pier and the connecting section on both sides are automatically measured at 8:00, 14:00 and 20:00 every day to obtain millimeter-level relative displacement and monitor the overall deformation trend of the gate body (such as uneven settlement on both sides and tilt of the gate body).

[0072] Laser displacement meter microscopic capture:

[0073] Laser displacement meters (range ≥ 500mm, resolution ≤ 0.01mm) are deployed at the same monitoring point. Through the principle of laser beam reflection time difference (or phase difference), the horizontal displacement, vertical settlement and surface relative displacement of the monitoring point are measured in real time, to make up for the shortcomings of the total station in monitoring high-frequency micro-deformation (such as thermal expansion and contraction caused by temperature changes), and form a deformation monitoring system of "macro trends + micro details".

[0074] 3. Multi-point tracking of leaked data

[0075] Deploy leakage pressure sensors or flow sensors downstream of the anti-seepage curtain at the gate foundation (to monitor the risk of anti-seepage structure failure), at the drainage holes (to monitor the seepage field pressure), and around the gate waterstop (to monitor the sealing of the waterstop device):

[0076] The pressure sensor senses the seepage water pressure through the pressure-sensitive element, combines the Bernoulli equation to reversely infer the seepage field distribution, and identifies the weak areas of the anti-seepage curtain;

[0077] Flow sensors use electromagnetic or ultrasonic types to directly measure the leakage volume per unit time. The two work together to locate the leakage source and quantify the leakage extent.

[0078] 4. Surface Defect Polarization Imaging

[0079] Polarized light source and imaging mechanism:

[0080] A linearly polarized light source (at four angles, 0°, 45°, 90°, and 135°) was used to uniformly illuminate the concrete surface of the sluice. Utilizing the difference in the Mueller matrix between the polarized light reflected by defects (cracks and honeycombed surfaces) and the intact surface (the polarization state of the reflected light at the defects rotates or depolarizes), a 20-megapixel polarization camera was used to synchronously capture four images with different polarization states.

[0081] Snake scanning strategy:

[0082] Carry out full coverage scanning of the sluice gate surface every week, with a single scan covering an area of ​​≥200m 2 It adopts a serpentine path to move back and forth, and the overlap rate of adjacent scanning strips is ≥30%, ensuring that there is no monitoring blind spot. It cooperates with high-precision displacement guide rails (positioning accuracy ±0.5mm) to achieve image stitching and coordinate calibration.

[0083] 2. Heterogeneous Data Transmission and Spatiotemporal Synchronization

[0084] 1. Multi-protocol fusion transmission architecture

[0085] Sensor data link:

[0086] Real-time data such as stress, deformation, and leakage are packaged into JSON format via industrial Ethernet (the data structure is:

[0087] `{"timestamp":"2025 04

[0089] 2508:00:00","sensor_id":"S01","data_value":1

[0090] 23.45,"checksum":"ABC123"}`), and is transmitted to the data processing server via a VPN encrypted channel to ensure data integrity; the transmission protocol supports Modbus / TCP (common protocol for industrial equipment) and OPCUA (cross-platform interoperability protocol), and is compatible with sensors from different manufacturers.

[0091] Image data link:

[0092] Polarization images are uploaded separately via the FTP protocol, and a resumable upload mechanism is used to ensure that the transmission time of a single image (≥20MB) is ≤15s to avoid occupying real-time data bandwidth. Lightweight processing (such as JPEG2000 compression with a compression ratio of 10:1) is performed before transmission to preserve polarization information while reducing the transmission load.

[0093] 2. Space-time synchronization mechanism

[0094] All equipment is connected to the Beidou / GPS timing module (timing accuracy ≤ 100ns), and the system clock is calibrated using the NTP protocol to ensure that the timestamp error of stress (10Hz), deformation (3 times a day), leakage (real-time) data and polarization images (weekly scans) is ≤ 10ms. At the same time, the sensor geographic coordinates are recorded during data collection (using a fixed control point coordinate system deployed at the sluice gate) to achieve spatiotemporal alignment of the monitoring data, laying the foundation for multi-parameter coupling analysis.

[0095] 3. Principles of multimodal data processing (noise filtering and feature extraction)

[0096] 1. Sensor data preprocessing pipeline

[0097] Kalman filter dynamic noise reduction:

[0098] For the dynamic noise (such as environmental vibration and temperature drift) in stress, deformation and leakage data, the extended Kalman filter algorithm is used to calculate the dynamic noise (such as environmental vibration and temperature drift) in the stress, deformation and leakage data. k|k-1 =F k X k-1|k-1 +B k u k Predict the state at time k, observation equation Z k =H k X k|k-1 +W k Fusion of measured values, iterative update of optimal estimate X k|k =X k|k-1 +K k (Z k -H k X k|k-1 ), filter out high-frequency noise and retain trend signals.

[0099] Outlier removal and data repair:

[0100] The Z-score method was used to calculate the mean μ and standard deviation σ of the data series, and outliers with absolute values ​​exceeding μ±3σ were eliminated. For missing data (such as fragment loss caused by communication interruption), the cubic spline interpolation algorithm was used, and piecewise cubic polynomial fitting was performed to ensure that the missing data completion accuracy within 72 consecutive hours was ≥98%.

[0101] 2. Full process of polarization image analysis

[0102] Preprocessing stage:

[0103] 3×3 median filter: Iterates over each pixel in the image and replaces the current value with the median value in the neighborhood to remove salt and pepper noise (such as sensor noise and surface stains and reflections);

[0104] Adaptive Histogram Equalization (CLAHE): Divides the image into 8×8 sub-blocks and performs independent histogram equalization on each sub-block to enhance the contrast between cracks and background and avoid overexposure caused by global equalization.

[0105] Defect segmentation and feature extraction:

[0106] Otsu threshold automatic segmentation: Calculate the inter-class variance of the image grayscale histogram

[0107]

[0108] , search for The maximum threshold T is used to binarize the image into the defect area (g(x,y)=1) and the background area (g(x,y)=0);

[0109] Morphological operation denoising: Remove isolated noise points with an area less than 100 pixels through dilation and erosion operations, retaining the true defect outline;

[0110] Canny edge detection: Dual thresholds (high threshold 0.3 × maximum gradient value, low threshold 0.1 × maximum gradient value) are used to detect crack edges. Hough transform is used to fit straight lines / curves to extract parameters such as crack length, width (pixel-level accuracy, 1 pixel ≈ 0.1 mm), and direction (angular resolution 1°). Euclidean distance matching is then performed with the preset defect library (matching error ≤ 5%).

[0111] 4. Inversion Principle of Multi-physics Field Coupling Model

[0112] 1. 3D finite element model construction details

[0113] Geometric modeling and meshing:

[0114] Based on the sluice gate design drawings (CAD file import), a 1:1 3D solid model was created in ANSYS Mechanical, including key components such as piers, gates, and foundations.

[0115] Tetrahedral grid division is adopted, with unit size ≤50cm (encrypted to 20cm in stress concentration areas such as the connection between piers and gates), and the total number of grids ≥500,000 to ensure calculation accuracy.

[0116] Physical meaning of coupling equations:

[0117] Mechanical equilibrium equation:

[0118] Describes the equilibrium state of the stress tensor σ under the gravity field ρg, which is used to solve the internal stress distribution of the structure;

[0119] Deformation constitutive equation:

[0120] δ = Dσ + ε0, which relates stress and deformation through the flexibility matrix D (a function of the material's elastic parameters), taking into account the initial strain ε0 of the concrete (such as shrinkage creep);

[0121] Seepage control equation:

[0122] Describing the seepage velocity q and the head gradient based on Darcy's law The relationship between the permeability coefficient k is used to simulate the leakage path.

[0123] 2. Data-driven model calibration process

[0124] Parameter initialization:

[0125] Input concrete elastic modulus (initial value 30GPa), Poisson's ratio (0.2), permeability coefficient (1×10 -6 cm / s) and other material parameters, and the boundary conditions are set in combination with the operating conditions of the sluice (such as upstream and downstream water levels and gate opening and closing loads).

[0126] Iterative optimization algorithm:

[0127] The Levenberg-Marquardt algorithm (LM algorithm) is used to minimize the root mean square error (RMSE) between the measured value and the model predicted value. The iterative formula is:

[0128] (J T J+λdiag(J T J))Δx=J T (yf(x))

[0129] Where J is the Jacobian matrix, λ is the damping parameter, and the weights of the Gauss-Newton method and the gradient descent method are dynamically adjusted to avoid local minima;

[0130] The termination condition was set as RMSE ≤ 0.5% or the number of iterations ≥ 1000 times, ensuring that the calibration error of the elastic modulus was ≤ 5% and the error of the permeability coefficient was ≤ 10%.

[0131] Real-time inversion applications:

[0132] The pre-processed historical data for more than five years (including more than 10,000 sets of data on stress, deformation, leakage, water level, etc.) is used as the training set. The model parameters are updated every hour to invert the stress distribution cloud map (accuracy ±2%), deformation displacement vector (resolution 0.01mm) and leakage velocity field (error ≤8%) of each unit of the current sluice, and quantify the coupled evolution process of "stress concentration → crack expansion → leakage channel formation".

[0133] 5. Graded warning and defect location mechanism

[0134] 1. Three-level threshold linkage warning logic

[0135] Threshold system design:

[0136] Level 1 warning (yellow): When a single parameter exceeds 80% of the threshold (e.g., stress ≥ design value × 0.8, deformation rate ≥ 1.6 mm / month, leakage flow ≥ 4 L / s), a local sound and light alarm is triggered, and the abnormal measurement point is highlighted on the monitoring platform;

[0137] Level 2 warning (orange): If any two parameters exceed the standard or a single parameter is abnormal for three consecutive times, encrypted monitoring will be activated (the deformation monitoring frequency is increased to once per hour, and the polarization imaging scanning cycle is shortened to three days). At the same time, a warning text message (including the abnormal parameters and measurement point locations) will be sent to the management department.

[0138] Level 3 warning (red): If three parameters exceed the standard at the same time, the leakage flow rate suddenly increases by ≥50%, or a structural defect is detected (crack width ≥0.2mm and depth ≥5cm), a global alarm will be triggered immediately, the relevant gate operations will be stopped, the defect position will be located through the GIS system, and a maintenance work order with repair suggestions (including defect image, geometric parameters, and impact analysis) will be automatically generated and pushed to the mobile phone app (supports iOS / Android) within 10 seconds.

[0139] 2. GIS visualization and defect location

[0140] The 3D model of the sluice gate is integrated with the electronic map, and the location of each monitoring point is displayed in real time on the GIS interface (red dots for stress sensors, yellow triangles for deformation monitoring, and blue squares for leakage detection). The warning status is dynamically indicated by flashing colors (yellow / orange / red);

[0141] Combined with the results of polarization image analysis, the crack coordinates (geocoding based on the scanning path) are mapped to the GIS model. Click on the defect annotation to view detailed parameters (for example, the crack is located at the No. 2-3 joint on the upstream surface of the pier, with a width of 0.3mm and a length of 65cm), realizing the closed-loop management of "early warning-positioning-disposal".

[0142] 6. System Collaboration Workflow

[0143] 1. Data collection layer:

[0144] The stress sensor collects data at a high frequency of 10 Hz → the deformation monitoring device is triggered three times a day → the leakage sensor collects real-time flow data → the polarization imaging module performs serpentine scanning every week to generate a multi-dimensional data set aligned in time and space (the average daily data volume is ≈ 5GB, including more than 1,000 polarization images).

[0145] 2. Transport layer:

[0146] Sensor data is packaged into JSON via industrial Ethernet and transmitted to the edge computing node through a fiber optic switch (supporting ring network redundancy and failover time ≤ 20ms), where invalid data is initially filtered out.

[0147] Polarization images are uploaded to the cloud server via FTP, stored in a distributed file system (HDFS), and indexed by timestamp and monitoring point ID.

[0148] 3. Processing layer:

[0149] The sensor data was processed through Kalman filtering, Z-score elimination, and cubic spline interpolation to generate a standard time series (1-minute resolution);

[0150] After denoising, enhancement, and segmentation, the polarization image outputs the defect coordinates and geometric parameters, which are associated with the sensor data through timestamps to form a four-tuple data set of "stress-deformation-leakage-image".

[0151] 4. Model layer:

[0152] The 3D finite element model loads the processed data in real time, and the material parameters are dynamically calibrated through the LM algorithm. The inversion results are output in the form of a grid cloud map, and the stress value, deformation variable, and leakage amount of each monitoring point are also output (error ≤ 5%).

[0153] 5. Application layer:

[0154] The early warning module compares the inversion results with the threshold in real time and triggers graded early warnings; the GIS system synchronously updates the status of monitoring points, and defect location information drives the generation of maintenance work orders, forming an intelligent closed loop of "monitoring-analysis-decision-making-feedback".

[0155] Based on the above principles, the system achieves full-scale coverage from millimeter-level surface defect detection to meter-level overall structural safety assessment, breaking through the limitations of isolated parameter analysis of traditional monitoring, and constructing a dynamic assessment system for sluice safety based on multi-physical field coupling, providing a technical paradigm of interdisciplinary integration for the structural health management of water conservancy projects.

[0156] The above description is merely a preferred embodiment of the present invention; however, the scope of protection of the present invention is not limited thereto. Any person skilled in the art who, within the technical scope disclosed by the present invention, makes equivalent substitutions or modifications based on the technical solutions and improved concepts of the present invention shall be covered by the scope of protection of the present invention.

Claims

1. A sluice structure safety monitoring system, characterized in that: It includes data acquisition module, data transmission module, data processing module, model building module, early warning module and polarization imaging module; The data acquisition module is used to collect stress data, deformation data, leakage data and surface image data of the sluice, and specifically includes: Stress sensors, using resistance strain sensors or vibrating wire sensors, are deployed at key stress-bearing locations including but not limited to sluice piers and gate support beams to collect stress and strain signals in real time; The deformation monitoring device includes a total station and a laser displacement meter. The total station is set at the connecting section of the two banks to monitor the relative displacement data of the top of the sluice body, the bottom of the pier and the connecting section of the two banks. The laser displacement meter is set at the top of the sluice body, the bottom of the pier and the connecting section of the two banks to monitor the horizontal displacement, vertical settlement and relative displacement data of the structure surface. Leakage detection devices, using leakage pressure sensors or flow sensors, are placed downstream of the anti-seepage curtain at the gate foundation, around the drainage holes and gate waterstops to monitor seepage pressure and leakage flow; The polarization imaging module includes: A polarized light source, using a linearly polarized light source or a circularly polarized light source, emits incident light in a specific polarization state, wherein the incident light in the specific polarization state is linearly polarized light of 0°, 45°, 90°, and 135°, and evenly illuminates the concrete surface of the sluice gate; Polarization camera, equipped with a multi-channel polarization filter, capable of simultaneously capturing images in at least two different polarization states, with a pixel resolution of no less than 20 megapixels; The data transmission module adopts an industrial-grade Ethernet or 4G / 5G wireless transmission module, is connected to the data acquisition module and the polarization imaging module, and transmits the stress data, deformation data, leakage data and polarization image data to the data processing module in real time through the TCP / IP protocol or the MQTT protocol; The data processing module includes a digital signal processing unit and an image analysis unit. The digital signal processing unit uses a Kalman filter algorithm to filter the received data. The Kalman filter algorithm equation is: X k|k =X k|k-1 +K k (Z k -H k X k|k-1 ) Among them, X k|k is the optimal estimated state at time k, X k|k-1 is the predicted state at time k based on time k_1, K k is the Kalman gain, Z k is the measured value at time k, H k For the measurement matrix, the cubic spline interpolation algorithm is used to fill in the missing data, and the Z-score method is used to identify and eliminate abnormal data outside 3σ; The image analysis unit sequentially performs median filtering noise reduction and histogram equalization enhancement processing on the polarization image, and uses the Otsu algorithm to automatically determine the threshold T for threshold segmentation. The threshold segmentation formula is: Among them, g(x,y) is the segmented image, f(x,y) is the original image, and then the defect area is extracted. The characteristic parameters of crack length, width, and direction are obtained through the Canny edge detection algorithm, and the Euclidean distance matching analysis is performed with the preset defect feature library; The model building module establishes a three-dimensional finite element model containing the stress-deformation-leakage dynamic coupling relationship based on the sluice structure mechanics model and seepage field theory. The coupling equation of the three-dimensional finite element model is: Among them, σ is the stress tensor, ρ is the material density, g is the acceleration of gravity, D is the flexibility matrix, ε0 is the initial strain, k is the permeability coefficient, and h is the head function. Then, using historical monitoring data for more than five years, the least squares method is used to calibrate the model material parameters and invert the real-time stress distribution, deformation displacement and leakage path of the sluice. The early warning module presets three levels of safety thresholds and compares the model inversion results with the thresholds in real time. When the stress exceeds 80% of the design value, the deformation rate is greater than 2mm / month, or the leakage flow rate is greater than 5L / s, a graded early warning signal is issued through an audible and visual alarm and a remote monitoring platform. At the same time, combined with the results of polarization image analysis, structural defects with crack widths greater than 0.2mm or depths greater than 5cm are located and marked.

2. A sluice structure safety monitoring system according to claim 1, characterized in that: The total station adopts a prism-free distance measurement mode with a distance measurement accuracy of ≤±(2mm+2ppm×D); the laser displacement meter has a measurement range of ≥500mm and a resolution of ≤0.01mm.

3. A sluice structure safety monitoring system according to claim 1, characterized in that: The data transmission module is equipped with an industrial-grade fiber optic switch, supports Modbus / TCP and OPC UA industrial protocols, has a data transmission delay of ≤50ms, and a bit error rate of ≤10-9.

4. A sluice structure safety monitoring system according to claim 1, characterized in that: The model building module adopts the ANSYS Mechanical and COMSOL Multiphysics joint simulation platform and realizes the parametric modeling of the stress-deformation-seepage coupling field through the APDL language.

5. A sluice structure safety monitoring system according to claim 1, characterized in that: The early warning module integrates the GIS geographic information system, displays the location and early warning status of each monitoring point of the sluice in real time, supports remote push of early warning information on the mobile phone APP, and the response time is ≤10s.

6. A method for monitoring the safety of a sluice structure, comprising applying a sluice structure safety monitoring system according to any one of claims 1 to 5, characterized in that: The following steps are involved: S1. Data acquisition: The stress sensor collects stress and strain signals at a frequency of 10 Hz. The deformation monitoring device automatically measures three times a day at 8:00, 14:00, and 20:

00. The leakage sensor monitors the seepage pressure and flow in real time. The polarization imaging module performs a full-coverage scan of the sluice surface every week, with a single scan covering an area ≥ 200 m 2 , collect images of 4 linear polarization states, and the incident light angles of the 4 linear polarization states are 0°, 45°, 90° and 135° respectively; S2. Data transmission: The sensor data is packaged into JSON format via industrial Ethernet. The JSON format data includes a timestamp, sensor ID, data value, and checksum. The data is then transmitted to the data processing server via a VPN encrypted channel. The polarization image is uploaded via the FTP protocol. The transmission time for a single image is ≤15s. S3. Data processing: The sensor data is sequentially subjected to Kalman filtering, Z-score outlier removal, and cubic spline interpolation. The polarization image is then subjected to 3×3 median filtering for noise reduction and adaptive histogram equalization for enhancement. The Otsu algorithm is used to calculate the optimal threshold T, and the defect area is segmented. Morphological operations are used to remove noise areas with an area of ​​less than 100 pixels, and the crack geometric features are extracted. S4. Model inversion: A three-dimensional geometric model is established based on the sluice gate design drawings. A tetrahedral mesh is formed, with a cell size of ≤50 cm. Initial material parameter values ​​are then input, including the concrete elastic modulus, Poisson's ratio, and permeability coefficient. Preprocessed historical data for more than five years is used as a training set. The model parameters are optimized using the Levenberg-Marquardt algorithm. Real-time stress values, deformation, and leakage at each monitoring point of the sluice gate are then inverted. S5. Warning judgment: When the inverse stress value ≥ safety threshold, deformation ≥ displacement threshold, or leakage ≥ flow threshold, a warning of the corresponding level is triggered. The safety threshold is the design stress × 0.8, the displacement threshold is 20 mm, and the flow threshold is 5 L / s. When the crack width is ≥0.2mm and the length is ≥50cm, it is marked as a structural defect. A maintenance work order is generated based on the location information, and the early warning information and defect report are simultaneously pushed to the sluice management department.

7. A method for monitoring the safety of a sluice structure according to claim 6, characterized in that: During model training in step S4, the iteration termination condition is set to root mean square error ≤ 0.5% or the number of iterations ≥ 1000 times, and the parameter calibration accuracy requires elastic modulus error ≤ 5% and permeability coefficient error ≤ 10%.

8. A method for monitoring the safety of a sluice structure according to claim 6, characterized in that: In step S1, the scanning path of the polarization imaging module adopts a serpentine coverage strategy, and the overlap rate of adjacent scanning strips is ≥30%, ensuring that surface defects are monitored without blind spots.

Images