Method for predicting life bearing capacity degradation of port and wharf caisson structure

Abstract

The invention discloses a life bearing capacity degradation prediction method for a port and wharf caisson structure, which comprises a data acquisition module, a coupling analysis processor and a dynamic early warning output module which are arranged, and is characterized in that the data acquisition module comprises a machine vision unit, a distributed optical fiber sensing unit and a parameterized seismic oscillation input module which are arranged; through cooperative work of multiple modules, accurate prediction of the caisson structure is realized. According to the life bearing capacity degradation prediction method of the port and wharf caisson structure, machine vision and optical fiber sensing are combined, surface-interior synchronous three-dimensional quantification of caisson corrosion damage is achieved, the problem of data discreteness of a traditional method is solved, physical mechanism constraints are provided through a finite element model, model errors are corrected through machine learning, and the service life bearing capacity degradation prediction method is achieved. The bearing capacity prediction precision under near-fault pulse type seismic oscillation is improved, a lightweight parallel computing architecture is designed, dynamic coupling calculation of corrosion monitoring data and seismic oscillation input is supported, and minute-level bearing capacity threshold value updating is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of civil engineering structure health monitoring and disaster prevention and mitigation technology, and in particular to a method for predicting the life bearing capacity degradation of a port terminal caisson structure. Background Art

[0002] Since the caisson structure is in the marine environment and is affected by a variety of complex factors, the performance of the caisson structure will be degraded. In particular, the caisson structure is easily damaged under the coupling of the corrosive environment and earthquake disasters, which reduces its safety. Therefore, it is necessary to predict the life bearing capacity degradation of the port terminal caisson structure to ensure the long-term safety, reliability and economy of the caisson structure. However, the existing port terminal caisson structure prediction system still has certain defects when used.

[0003] During use, traditional methods mostly analyze corrosion or earthquake effects independently, ignoring the coupling effect between the two, that is, corrosion weakens the cross-section of the caisson structure. When the earthquake inertia force increases, it will accelerate the destruction of the caisson structure. In addition, corrosion detection relies on manual visual inspection or local sensors, making it difficult to achieve three-dimensional continuous quantification of the depth and distribution of rust. The seismic input is insufficiently parameterized, and there is a lack of targeted modeling of near-fault pulse seismic motion, making data collection discrete. In addition, existing bearing capacity prediction models are mostly based on empirical formulas and fail to couple multi-scale physical mechanisms such as material nonlinearity, geometric damage and contact interface slip, resulting in insufficient model accuracy.

[0004] In response to the above problems, it is urgently necessary to carry out innovative design based on the original life bearing capacity degradation prediction method of the port terminal caisson structure. Therefore, we proposed a life bearing capacity degradation prediction method for the port terminal caisson structure that can well solve the above problems. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for predicting the life-span bearing capacity degradation of a port terminal caisson structure, so as to solve the problems raised in the above-mentioned background technology. The traditional methods on the market mostly analyze corrosion or seismic effects independently, ignoring the coupling effect between the two, that is, corrosion weakens the cross-section of the caisson structure. When the seismic inertia force increases, it will accelerate the destruction of the caisson structure. In addition, corrosion detection relies on manual visual inspection or local sensors, making it difficult to achieve three-dimensional continuous quantification of the depth and distribution of rust. The seismic input parameterization is insufficient, and there is a lack of targeted modeling of near-fault pulse seismic motion, which makes data acquisition discrete. In addition, the existing bearing capacity prediction models are mostly based on empirical formulas, which fail to couple multi-scale physical mechanisms such as material nonlinearity, geometric damage and contact interface slip, resulting in insufficient model accuracy.

[0006] To achieve the above-mentioned objectives, the present invention provides the following technical solutions: a method for predicting the life-span bearing capacity degradation of a port terminal caisson structure, comprising a data acquisition module, a coupling analysis processor, and a dynamic warning output module, wherein the data acquisition module comprises a machine vision unit, a distributed fiber optic sensing unit, and a parameterized seismic input module. Through the collaborative operation of multiple modules, an accurate prediction of the bearing capacity degradation of the caisson structure under the coupling of corrosion and earthquake disasters can be achieved.

[0007] Preferably, the coupling analysis processor includes a set corrosion-section weakening model, a multi-hazard coupling dynamic model and a machine learning correction module.

[0008] Preferably, the coupling analysis processor solves the structural response under the action of corrosion-seismic coupling through an explicit dynamic analysis method, and uses a random forest algorithm to correct the errors of the simulation results. The input features of the random forest algorithm include corrosion depth, seismic PGA, and section weakening coefficient. The corrected error threshold is ≤10%.

[0009] Preferably, the dynamic warning output module includes calculation of the set bearing capacity threshold and generation of warning signals, and dynamically updates the remaining bearing capacity threshold based on the real-time damage index and reliability index. When the bearing capacity degradation rate exceeds the preset safety threshold, an audible and visual alarm is triggered and maintenance decision recommendations are pushed.

[0010] Preferably, the hardware deployment of the life bearing capacity degradation prediction method of the caisson structure includes: a machine vision unit, a fiber optic sensing unit and an edge computing terminal. The machine vision unit includes an explosion-proof industrial camera and a laser scanner installed on the top plate and side wall of the caisson through a track-type mobile bracket. The fiber optic sensor spacing is ≤0.5m, and the sampling frequency is ≥100Hz.

[0011] Preferably, the machine vision unit uses a U-Net network to achieve semantic segmentation of the corroded area, and fuses it with the laser point cloud data to generate a three-dimensional section weakening coefficient matrix.

[0012] Preferably, the distributed fiber optic sensing unit is provided with fiber optic Bragg grating sensors arranged along the vertical and horizontal main reinforcements of the caisson, and the edge computing terminal is equipped with an NVIDIA Jetson AGX Xavier processor and a built-in coupling analysis algorithm to support on-site real-time data processing.

[0013] Preferably, the software process of the life bearing capacity degradation prediction method of the caisson structure includes: data preprocessing, coupling analysis and early warning decision-making, wherein the data preprocessing is carried out by dehazing and edge enhancement of the corrosion image, outputting a pixel-level corrosion mask through a deep learning model, and fusing the optical fiber strain data and the corrosion distribution to construct a three-dimensional section weakening coefficient matrix.

[0014] Preferably, the coupling analysis is performed by invoking the OpenSees platform to establish a nonlinear finite element model of the caisson taking into account corrosion damage, inputting near-fault ground motion, and performing dynamic time history analysis.

[0015] Compared with the existing technology, the beneficial effects of the present invention are as follows: the life-span bearing capacity degradation prediction method of the port terminal caisson structure combines machine vision and fiber optic sensing to achieve "surface-interior" synchronous three-dimensional quantification of caisson corrosion damage, solves the data discreteness problem of traditional methods, provides physical mechanism constraints through finite element models, uses machine learning to correct model errors, improves the bearing capacity prediction accuracy under near-fault pulse-type earthquake motion, designs a lightweight parallel computing architecture, supports dynamic coupling calculation of corrosion monitoring data and earthquake motion input, and realizes minute-level bearing capacity threshold updates. The specific contents are as follows: (1) Combining machine vision and fiber optic sensing, the "surface-interior" synchronous three-dimensional quantification of caisson corrosion damage is achieved, solving the data discreteness problem of traditional methods. The finite element model provides physical mechanism constraints, and machine learning is used to correct model errors, thereby improving the bearing capacity prediction accuracy under near-fault pulse earthquake motion. A lightweight parallel computing architecture is designed to support the dynamic coupling calculation of corrosion monitoring data and earthquake motion input, and to achieve minute-level bearing capacity threshold updates. Compared with traditional empirical formulas, the bearing capacity prediction error is reduced by more than 40%, supporting online monitoring and minute-level analysis, meeting the needs of engineering emergency response, and being suitable for the full life cycle management of caisson structures in complex marine environments.

[0016] (2) Select the NVIDIA Jetson AGX Xavier processor as the core component of the edge computing terminal. The processor has powerful computing capabilities and can meet the needs of on-site real-time data processing. Build a terminal hardware platform, including installing necessary storage devices and power modules, and correctly connect the processor with the storage device and data acquisition interface to ensure that data can be smoothly transmitted to the processor for processing. Install the coupling analysis algorithm program in the terminal device. The algorithm program should undergo rigorous testing and optimization to ensure that it can accurately and efficiently process various types of monitoring data in actual operation.

[0017] (3) After acquiring the rust image of the caisson surface from the machine vision unit, the image is first dehazed using an image dehazing algorithm to eliminate image blur caused by factors such as fog in the port environment and enhance image clarity and contrast. Common dehazing algorithms such as the dark channel prior dehazing algorithm can effectively achieve this goal.

[0018] (4) According to the fault type and focal mechanism of the site, the near-fault seismic motion parameterized model is used to generate seismic motion time history data with velocity pulse characteristics. The peak acceleration (PGA) of the input seismic motion is set to 0.3g, and the velocity pulse period (T) is set to 2s. These are typical seismic motion parameters determined based on the seismic hazard analysis of the port area and the engineering design requirements. The generated seismic motion time history data are input into the established caisson finite element model considering corrosion damage, and the model is subjected to dynamic time history analysis using the explicit dynamic analysis method.

[0019] (5) Based on the damage index calculated in real time and the pre-set reliability index (β ≥ 2.0), the residual bearing capacity threshold of the caisson structure is dynamically updated using structural reliability analysis theory and related calculation formulas. As the monitoring data is continuously updated and the damage to the structure develops under the influence of corrosion and earthquakes, the residual bearing capacity threshold will change in real time to reflect the current actual bearing capacity status of the structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a schematic diagram of the wharf caisson structure prediction system of the present invention; Figure 2 This is a schematic diagram of the data acquisition module of the present invention; Figure 3 This is a schematic diagram of the coupling analysis processor of the present invention; Figure 4 This is a schematic diagram of the dynamic warning output module of the present invention; Figure 5 This is a schematic diagram of the core functions of the present invention; Figure 6 Schematic diagram of a specific embodiment of the present invention; Figure 7 This is a schematic diagram of the hardware deployment of the present invention; Figure 8 Schematic diagram of the software flow of the present invention; Figure 9 This is the early warning decision logic block diagram of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] Example: In this example, by combining machine vision and fiber optic sensing, the "surface-interior" synchronous three-dimensional quantification of caisson corrosion damage is achieved, solving the data discreteness problem of traditional methods. The finite element model provides physical mechanism constraints, and machine learning is used to correct model errors. The bearing capacity prediction accuracy under near-fault pulse earthquake motion is improved. A lightweight parallel computing architecture is designed to support the dynamic coupling calculation of corrosion monitoring data and earthquake motion input, and to achieve minute-level bearing capacity threshold updates, such as Figures 1-9 The technical solution shown in the figure includes a data acquisition module, a coupling analysis processor and a dynamic warning output module, wherein the data acquisition module includes a machine vision unit, a distributed fiber optic sensing unit and a parameterized seismic input module. Through the collaborative work of multiple modules, the bearing capacity degradation of the caisson structure under the coupling of corrosion and earthquake disasters can be accurately predicted. The coupling analysis processor includes a corrosion-section weakening model, a multi-hazard coupling dynamic model and a machine learning correction module. The coupling analysis processor solves the structural response under the coupling of corrosion and earthquake through an explicit dynamic analysis method, and uses a random forest algorithm to correct the error of the simulation results. The input features of the random forest algorithm include corrosion depth, seismic PGA, and section weakening coefficient. The corrected error threshold is ≤10%. The dynamic warning output module includes a bearing capacity Threshold calculation and early warning signal generation, dynamically update the remaining bearing capacity threshold according to the real-time damage index and reliability index, when the bearing capacity degradation rate exceeds the preset safety threshold, trigger the sound and light alarm and push maintenance decision suggestions, the hardware deployment of the caisson structure life bearing capacity degradation prediction method includes: machine vision unit, fiber optic sensing unit and edge computing terminal, the machine vision unit includes an explosion-proof industrial camera and a laser scanner installed on the top plate and side wall of the caisson through a track-type mobile bracket, the fiber optic sensor spacing is ≤0.5m, the sampling frequency is ≥100Hz, the machine vision unit uses the U-Net network to realize the semantic segmentation of the rusted area, and fuses it with the laser point cloud data to generate a three-dimensional section weakening coefficient matrix, the distributed fiber optic sensing unit is equipped with fiber grating sensors arranged along the vertical and horizontal main reinforcement of the caisson, and the edge computing terminal is equipped with NVIDIA The Jetson AGX Xavier processor has a built-in coupling analysis algorithm that supports real-time on-site data processing. The software process for predicting the life bearing capacity degradation of caisson structures includes data preprocessing, coupling analysis, and early warning decision-making. Data preprocessing involves dehazing and edge enhancement of corrosion images, outputting pixel-level corrosion masks through a deep learning model, and fusing fiber strain data with corrosion distribution to construct a three-dimensional cross-sectional weakening coefficient matrix. The coupling analysis uses the OpenSees platform to establish a nonlinear finite element model of the caisson that takes corrosion damage into account. Near-fault ground motions are then input for dynamic time-history analysis.

[0023] In hardware deployment, first, the machine vision unit needs to be installed. It is necessary to select a suitable explosion-proof industrial camera and ensure that its resolution reaches or exceeds 20 million pixels. This can ensure that the acquired image of the caisson surface has sufficient clarity to accurately identify rusted areas. The industrial camera has an anti-corrosion coating standard that meets the IP68 waterproof grade standard, which can effectively resist erosion in harsh environments. In addition, camera fixing brackets are installed at key locations such as the four corners of the caisson top plate and the middle position of the side wall. The bracket should have strong stability and can be fixed to the caisson surface by welding or high-strength bolts. The industrial camera is mounted on the bracket and the camera's shooting angle is adjusted so that the camera can cover the largest possible caisson surface area. To achieve panoramic scanning, a track-type mobile bracket is installed around the caisson. The track is laid along the contour of the caisson to ensure that the industrial camera can move smoothly on the track and maintain the continuity of image acquisition during movement.

[0024] Secondly, during the laying of the fiber optic sensing unit, the fiber optic Bragg grating sensors need to be arranged along the vertical and horizontal main reinforcements during the caisson reinforcement binding process. A sensor is fixed on the vertical main reinforcement at intervals of no more than 0.5m, and the same is true for the horizontal ones to ensure effective monitoring of all parts of the caisson. A special fixing fixture is used to tightly fix the fiber optic Bragg grating sensor on the main reinforcement to avoid the sensor position from shifting during the concrete pouring process. The fiber optic sensor has excellent anti-electromagnetic interference capability and an isolation of 50dB. Ensure that the connecting optical fiber of the sensor is arranged along the reinforcement to avoid excessive bending or stretching of the optical fiber. Connect the optical fiber lines of all fiber optic Bragg grating sensors and aggregate them to the data acquisition terminal to ensure the connectivity of the entire fiber optic sensing network. Set the sampling frequency of the data acquisition terminal to 100Hz or above to ensure that subtle strain anomalies caused by concrete strain, crack expansion and internal reinforcement corrosion can be captured in a timely manner.

[0025] Finally, during the construction of the edge computing terminal, the NVIDIA Jetson AGX Xavier processor was selected as the core component of the edge computing terminal. This processor has powerful computing capabilities and can meet the needs of real-time data processing on site. The terminal hardware platform was built, including the installation of necessary storage devices and power modules, and the correct connection of the processor with the storage device and data acquisition interface to ensure that the data can be smoothly transmitted to the processor for processing. The coupling analysis algorithm program was installed in the terminal device. The algorithm program should undergo rigorous testing and optimization to ensure that it can accurately and efficiently process various monitoring data in actual operation. After long-term operation testing, the system has a failure-free rate of ≥99.9% for one year of continuous operation, demonstrating extremely high stability and reliability.

[0026] During software operation, after obtaining the rust image of the caisson surface from the machine vision unit, the image is first dehazed using an image dehazing algorithm to eliminate image blur caused by factors such as fog in the port environment and enhance image clarity and contrast. Commonly used dehazing algorithms such as the dark channel prior dehazing algorithm can effectively achieve this goal.

[0027] Next, the dehazed image is processed using an edge enhancement algorithm to highlight the edges of the corroded areas. For example, the Sobel operator and the Canny edge detection algorithm are used to sharpen the outlines of the corroded areas, facilitating subsequent analysis and identification. The dehazed and edge-enhanced image is then fed into a pre-trained deep learning model, such as a U-Net-based image semantic segmentation model. The training dataset consists of 1,000 images of corroded port caissons, covering various lighting conditions and corrosion levels. The model training parameters are set to a learning rate of 0.001 and 1,000 iterations, achieving a segmentation accuracy of ≥0.85. A random forest algorithm, with input features such as stress, displacement, and corrosion depth, and a set number of trees to 500, effectively reduces the prediction error from 15% to 8%. In addition, the window size of the dark channel dehazing algorithm is set to 15×15, and it takes less than 50ms to process a single frame of image. The model has been trained with a large number of rust image samples and can perform pixel-level analysis on the input image, outputting an accurate rust mask and identifying whether each pixel in the image belongs to the rust area.

[0028] Then, the concrete strain data and the corrosion distribution information obtained by the machine vision unit are obtained from the fiber optic sensing unit. The strain data monitored by the fiber optic sensors at different positions are combined with the corrosion conditions at the corresponding positions. According to the principles of concrete structural mechanics and the impact of corrosion on the performance of concrete and steel bars, the cross-sectional weakening coefficient caused by corrosion is calculated at each monitoring position. For example, the reduction ratio of the effective cross-sectional area of the steel bars is calculated according to the degree of steel bar corrosion, as well as the impact of the strength reduction of the concrete due to corrosion on the cross-sectional bearing capacity. The cross-sectional weakening coefficients of each monitoring position are arranged and combined according to the three-dimensional spatial coordinates to construct a three-dimensional cross-sectional weakening coefficient matrix. This matrix comprehensively reflects the changes in the cross-sectional bearing capacity of the caisson due to corrosion at different positions and depths, providing a key data basis for subsequent bearing capacity degradation analysis, greatly improving the accuracy and comprehensiveness of the assessment of the caisson corrosion condition.

[0029] Then, the professional structural analysis software OpenSees platform was called to establish a nonlinear finite element model of the caisson based on the actual geometric dimensions, material parameters and other information of the caisson. During the modeling process, the nonlinear mechanical properties of concrete and steel bars were fully considered, such as the plasticity and cracking behaviors of concrete and the yielding and strengthening characteristics of steel bars. The three-dimensional section weakening coefficient matrix obtained through data preprocessing was introduced into the finite element model to simulate the damage caused by corrosion to the caisson structure. By adjusting and optimizing the model parameters, the model can accurately reflect the changes in the mechanical properties of the actual caisson structure in a corrosive environment.

[0030] Then, according to the fault type and focal mechanism of the site, the near-fault seismic motion parameterized model was used to generate seismic motion time history data with velocity pulse characteristics. The peak acceleration (PGA) of the input seismic motion was set to 0.3g, and the velocity pulse period (T) was set to 2s. These are typical seismic motion parameters determined based on the seismic hazard analysis of the port area and engineering design requirements. The generated seismic motion time history data was input into the established caisson finite element model considering corrosion damage, and the model was subjected to dynamic time history analysis using the explicit dynamic analysis method. During the analysis, the model continuously updated the displacement, velocity and acceleration responses of the structure according to the input seismic motion, and solved the elastic-plastic time history response of the structure under earthquake action through numerical calculation. During the calculation process, the model was used to calculate the time history of the structure according to the depth of corrosion. The non-uniform geometric shape of the caisson cross-section is reconstructed based on the degree distribution, and the effective bearing area reduction coefficient and material strength degradation rate are calculated through relevant calculation formulas. At the same time, key parameters such as displacement, stress and damage index of the structure under earthquake action are output in real time. These parameters reflect the changes in the mechanical properties of the caisson structure under the coupling of corrosion and earthquake. The random forest algorithm is used to correct the errors of the numerical simulation results obtained by dynamic time history analysis. A large amount of sample data is extracted from historical monitoring data and laboratory scale test data, and the random forest algorithm model is trained to enable it to accurately identify the error characteristics in the numerical simulation results and correct the current simulation results, thereby improving the accuracy of the prediction results and providing a solid guarantee for the accurate assessment of bearing capacity degradation.

[0031] Finally, based on the damage index calculated in real time and the preset reliability index (β≥2.0), the structural reliability analysis theory and related calculation formulas are used to dynamically update the residual bearing capacity threshold of the caisson structure. With the continuous updating of monitoring data and the development of damage to the structure under the influence of corrosion and earthquakes, the residual bearing capacity threshold will change in real time to reflect the current actual bearing capacity status of the structure. The bearing capacity degradation rate ΔR / R0 is calculated in real time, where R0 is the initial bearing capacity of the caisson structure and ΔR is the bearing capacity loss due to corrosion and earthquakes. When the bearing capacity degradation rate ΔR / R0 reaches or exceeds 15%, the first-level warning is activated, and an alarm signal is issued through the system's sound and light alarm device to remind relevant staff to pay attention. At the same time, the system is based on the preset According to the decision-making rules, it is recommended to use a carbon fiber reinforcement scheme to locally reinforce the caisson structure to improve the bearing capacity of the structure; when the bearing capacity degradation rate ΔR / R0 reaches or exceeds 30%, the second-level warning is activated. At this time, the bearing capacity of the structure has seriously decreased, and the system will issue a stronger sound and light alarm signal and recommend immediate shutdown and maintenance. Relevant staff need to conduct a comprehensive inspection and assessment of the caisson structure and formulate a detailed overall repair plan to ensure the safety and normal use function of the caisson structure; if the bearing capacity degradation rate does not exceed 15%, the monitoring status of the caisson structure will be maintained continuously. According to the above hardware deployment and software process, data collection, preprocessing, coupling analysis and early warning decision-making are continuously performed in a cycle to grasp the life bearing capacity degradation of the caisson structure in real time.

[0032] After deploying a device for predicting the lifespan bearing capacity degradation of a port terminal caisson structure on a caisson at an offshore port, detection revealed 8mm corrosion on the caisson's sidewalls and a 12% cross-sectional weakening rate. Simulations, combined with near-fault ground motion (PGA = 0.4g), revealed excessive node bending moments. The system quickly triggered an early warning and recommended shotcrete reinforcement. Upon completion, the caisson's bearing capacity recovered to 92% of its design value.

[0033] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. 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 predicting the life bearing capacity degradation of a port terminal caisson structure, comprising a data acquisition module, a coupling analysis processor, and a dynamic early warning output module, characterized in that: The data acquisition module includes a machine vision unit, a distributed fiber optic sensing unit, and a parameterized seismic input module. Through the collaborative work of multiple modules, it can achieve accurate prediction of the bearing capacity degradation of the caisson structure under the coupling of corrosion and earthquake hazards.

2. The method for predicting the life bearing capacity degradation of a port terminal caisson structure according to claim 1, characterized in that: The coupling analysis processor includes a set corrosion-section weakening model, a multi-hazard coupling dynamic model and a machine learning correction module.

3. The method for predicting the life bearing capacity degradation of a port terminal caisson structure according to claim 2, characterized in that: The coupled analysis processor solves the structural response under the coupled effects of corrosion and earthquakes through an explicit dynamic analysis method and uses a random forest algorithm to correct the errors in the simulation results. The input features of the random forest algorithm include corrosion depth, ground motion PGA, and section weakening coefficient. The corrected error threshold is ≤10%.

4. The method for predicting the life bearing capacity degradation of a port terminal caisson structure according to claim 1, characterized in that: The dynamic warning output module includes the calculation of the set bearing capacity threshold and the generation of warning signals. According to the real-time damage index and reliability index, the remaining bearing capacity threshold is dynamically updated. When the bearing capacity degradation rate exceeds the preset safety threshold, an audible and visual alarm is triggered and maintenance decision recommendations are pushed.

5. The method for predicting the life bearing capacity degradation of a port terminal caisson structure according to claim 1, characterized in that: The hardware deployment of the life bearing capacity degradation prediction method of the caisson structure includes: a machine vision unit, a fiber optic sensing unit and an edge computing terminal. The machine vision unit includes an explosion-proof industrial camera and a laser scanner installed on the top plate and side wall of the caisson through a track-type mobile bracket. The fiber optic sensor spacing is ≤0.5m, and the sampling frequency is ≥100Hz.

6. The method for predicting the life bearing capacity degradation of a port terminal caisson structure according to claim 5, characterized in that: The machine vision unit uses a U-Net network to achieve semantic segmentation of the rusted area and fuses it with the laser point cloud data to generate a three-dimensional cross-section weakening coefficient matrix.

7. The method for predicting the life bearing capacity degradation of a port terminal caisson structure according to claim 5, characterized in that: The distributed fiber optic sensing unit is equipped with fiber optic Bragg grating sensors arranged along the vertical and horizontal main reinforcement of the caisson. The edge computing terminal is equipped with an NVIDIA Jetson AGX Xavier processor and a built-in coupling analysis algorithm to support real-time data processing on site.

8. The method for predicting the life bearing capacity degradation of a port terminal caisson structure according to claim 1, characterized in that: The software process of the life bearing capacity degradation prediction method of the caisson structure includes: data preprocessing, coupling analysis and early warning decision-making. Among them, the data preprocessing is to defog and enhance the edge of the corrosion image, output pixel-level corrosion mask through the deep learning model, and fuse the optical fiber strain data with the corrosion distribution to construct a three-dimensional section weakening coefficient matrix.

9. The method for predicting the life bearing capacity degradation of a port terminal caisson structure according to claim 8, characterized in that: The coupling analysis uses the OpenSees platform to establish a nonlinear finite element model of the caisson taking into account corrosion damage, inputs near-fault ground motion, and performs dynamic time history analysis.

Images