Chloride ion erosion dynamic imaging system and method based on multi-modal sensing fusion

Abstract

The invention discloses a chloride ion erosion dynamic imaging system and method based on multi-modal sensing fusion. The system comprises an embedded sensor array module used for realizing multi-parameter in-situ sensing of Cl concentration, pH value, corrosion current, strain and temperature; the multi-field coupling modeling module is used for constructing a Cl diffusion-steel bar corrosion-concrete cracking strong coupling model and outputting a corrosion expansion force distribution cloud picture; the dynamic inversion module is used for generating a continuous monitoring field through space-time coding; the three-dimensional visualization engine module is used for constructing a digital twin base and realizing dynamic rendering of a red-yellow-green three-color erosion thermodynamic diagram; and the mobile inspection interaction module is used for superposing the erosion thermodynamic diagram and the steel bar deformation data on the intelligent glasses, supporting a voice instruction to generate a structured inspection log and offline operation, and automatically pushing a maintenance work order containing a disposal priority and a material list based on the erosion risk level. The invention provides a technical benchmark for the durability management of the infrastructure.

Description

Technical Field

[0001] This invention relates to the field of durability monitoring technology for concrete structures in shield tunnels, and in particular to a dynamic imaging system and method for chloride ion erosion based on multimodal sensor fusion. Background Technology

[0002] In the maintenance of offshore tunnel segments, dynamic imaging and accurate prediction of chloride ion erosion processes can effectively slow down the rate of steel corrosion, reduce excessive maintenance costs, thereby extending structural life and optimizing operation and maintenance resources. Currently, chloride ion erosion monitoring of shield tunnel segments mainly relies on manual drilling for sampling or localized installation of single-point sensors, with assessment of erosion status through laboratory testing or periodic inspections. Traditional methods require interrupting construction to obtain concrete core samples and using empirical formulas to extrapolate the overall erosion trend. Chloride ion erosion assessment primarily depends on manual drilling for sampling, laboratory testing, or periodic inspections using single-point sensors. This not only requires damaging the segment structure to obtain concrete core samples, significantly increasing the risk of microcracks inducing accelerated erosion, but also leads to frequent instances of insufficient local repairs or over-maintenance, resulting in significant material waste due to experience-driven static assessments.

[0003] However, manual sampling in related technologies can damage structural integrity, and single-point sensors struggle to capture spatial variations. This not only increases the risk of segment damage (such as accelerated erosion due to borehole microcracks) but also fails to quantify the chain reaction of ion migration, corrosion, and cracking, leading to insufficient local repairs or over-maintenance. Especially in high-salt coastal environments, experience-driven static assessments are insufficient to predict hidden defects, exacerbating the risk of steel reinforcement rust expansion and cracking. Furthermore, monitoring blind spots may delay critical maintenance windows. Achieving dynamic, visual tracking of the erosion process and providing real-time guidance for precise maintenance to balance structural lifespan and economic costs has become a core challenge in improving the durability of marine tunnels.

[0004] In recent years, with the development of MEMS sensing, multiphysics simulation, and AR visualization technologies, especially the engineering applications of bidirectional LSTM prediction models, Monte Carlo confidence assessment, and lightweight BIM engines, new approaches have been provided for the dynamic inversion and closed-loop maintenance of erosion processes. Intelligent operation and maintenance of tunnel segments has become an industry trend. If embedded sensor arrays, multi-field coupled modeling, and data-driven visualization technologies can be integrated to build a unified "perception-prediction-decision" system, it will significantly improve the accuracy of early warning of hidden defects, reduce maintenance costs, and avoid secondary structural damage, thus promoting the evolution of tunnel operation and maintenance towards digitalization and intelligence. Summary of the Invention

[0005] This invention proposes a dynamic imaging system and method for chloride ion corrosion based on multimodal sensor fusion, which solves the problems of manual sampling destroying structural integrity, single-point monitoring making it difficult to quantify the ion migration-corrosion-cracking chain reaction, and insufficient visualization of hidden diseases in related technologies, and realizes closed-loop optimization of dynamic tracking of corrosion process and maintenance decision-making.

[0006] On the one hand, to achieve the above objectives, the present invention provides a dynamic imaging system for chloride ion erosion based on multimodal sensor fusion, comprising:

[0007] Embedded sensor array module: used to realize multi-parameter in-situ sensing of Cl⁻ concentration, pH value, corrosion current, strain and temperature, and to provide initial conditions and boundary constraints for physical equations for multi-field coupling modeling module;

[0008] Multi-field coupling modeling module: used to construct a strongly coupled model of Cl⁻ diffusion-steel corrosion-concrete cracking, extract boundary conditions based on crack topology recognition algorithm, and output corrosion expansion force distribution cloud map;

[0009] Dynamic inversion module: Based on the corrosion expansion force distribution cloud map, it generates a continuous monitoring field through spatiotemporal coding, uses bidirectional LSTM to predict the erosion depth in the future preset time period, and quantifies the prediction reliability based on the Monte Carlo Dropout method;

[0010] 3D visualization engine module: used to build a digital twin base, realize dynamic rendering of red-yellow-green erosion heat map, and provide quantitative indicators;

[0011] Mobile inspection interaction module: It is used to overlay erosion heat maps and steel deformation data on smart glasses through SLAM positioning and lightweight rendering engine. It supports voice command to generate structured inspection logs and offline operations, and automatically pushes maintenance work orders containing treatment priorities and material lists based on erosion risk level.

[0012] Preferably, the embedded sensor array module includes:

[0013] Multi-parameter sensing unit: used to deploy MEMS electrochemical sensor arrays according to the spacing of the pipe segment joints, capture Cl⁻ concentration, pH value and corrosion current data, and simultaneously deploy distributed optical fibers to measure strain and temperature data;

[0014] Pressure-resistant encapsulation unit: used to form a protective structure with a silicon carbide ceramic shell and a nano-modified epoxy resin filler layer;

[0015] Array calibration unit: used to dynamically eliminate electromagnetic interference from the tunnel boring machine, and combined with temperature compensation circuit to stabilize the fluctuation range of Cl⁻ concentration measurement within the preset range.

[0016] Preferably, the multi-field coupling modeling module includes:

[0017] Erosion kinetics unit: used to establish a strongly coupled model of the unsteady Cl⁻ diffusion equation and the steel electrochemical corrosion rate equation, and to introduce the concrete crack propagation phase field theory to quantify the influence of microcracks on the ion diffusion coefficient;

[0018] Boundary condition element: used to extract the boundary of the real crack network through the microcrack topology recognition algorithm, and dynamically correct the boundary value of surface ion concentration by accessing the salinity / humidity data of the weather station in real time;

[0019] Real-time solver unit: Used to decompose the concrete mesh model into multiple computation nodes using a GPU parallel finite element architecture. When the fiber optic sensor detects a sudden increase in strain, it automatically triggers recalculation and outputs a cloud map of the expansion force distribution of corrosion products.

[0020] Preferably, the dynamic inversion module includes:

[0021] Spatiotemporal coding unit: used to synchronize sensor network data and generate a continuous monitoring field of Cl⁻ concentration using the Kriging spatial interpolation algorithm;

[0022] LSTM prediction unit: It is used to output the erosion depth prediction for a future time period based on the spatiotemporal sequence of a preset time period input by a bidirectional LSTM neural network, and the absolute error of the prediction depth result does not exceed a preset threshold.

[0023] Confidence assessment unit: used to generate a prediction probability distribution map, and when the root mean square error of three consecutive predictions is greater than the preset threshold, it automatically loads the incremental dataset to start weight adaptive retraining, and the updated model is automatically deployed to the value prediction pipeline.

[0024] Preferably, the 3D visualization engine module includes:

[0025] BIM Fusion Unit: Used to analyze the geometric topology of Revit segment models, bind sensor spatial coordinates with steel reinforcement grid coordinates, and build a digital twin base;

[0026] Heatmap generation unit: used to dynamically render red-yellow-green erosion heatmaps based on the risk level shading engine, and to overlay historical erosion paths using transparency gradient technology;

[0027] Data-driven unit: used to automatically generate cross-sectional erosion depth diagrams of pipe segments and output quantitative indicators by combining the steel reinforcement rust expansion deformation simulation algorithm.

[0028] Preferably, the mobile inspection interaction module includes:

[0029] AR overlay unit: used to dynamically overlay erosion heatmaps onto smart glasses using a SLAM positioning system and a lightweight rendering engine;

[0030] Human-computer interaction unit: used to generate structured inspection logs in response to voice commands;

[0031] Operation and maintenance decision unit: used to push maintenance work orders according to the erosion risk level.

[0032] On the other hand, to achieve the above objectives, the present invention also provides a method for dynamic imaging of chloride ion erosion based on multimodal sensor fusion, applied to a dynamic imaging system for chloride ion erosion based on multimodal sensor fusion, comprising:

[0033] Data on Cl⁻ concentration, pH value, corrosion current, strain, and temperature are acquired through an embedded sensor array module, and measurement accuracy is ensured through pressure-resistant packaging and noise calibration.

[0034] A strongly coupled model of Cl⁻ diffusion-steel corrosion-concrete cracking is constructed using a multi-field coupling modeling module, and the boundary conditions are dynamically corrected and a corrosion expansion force cloud map is output.

[0035] Based on the corrosion expansion force cloud map, a continuous monitoring field is generated through the dynamic inversion module to predict the erosion depth within a preset time period in the future, and the model is retrained when the error exceeds the limit.

[0036] The 3D visualization engine module renders dynamic erosion heatmaps and outputs several quantitative indicators.

[0037] The mobile inspection interaction module presents the erosion status in an AR environment and automatically pushes maintenance work orders.

[0038] By optimizing model parameters based on prediction error feedback, a closed-loop optimization of erosion imaging and operation and maintenance decisions is achieved.

[0039] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a dynamic imaging method for chloride ion erosion based on multimodal sensor fusion.

[0040] A computer-readable storage medium storing a computer program that, when executed by a processor, implements a dynamic imaging method for chloride ion erosion based on multimodal sensor fusion.

[0041] Compared with the prior art, the present invention has the following advantages and technical effects:

[0042] (1) Significant improvement in measurement accuracy: The environmental noise learning algorithm reduces the fluctuation range of Cl⁻ concentration measurement from ±5%, improving spatial consistency;

[0043] (2) Enhanced prediction reliability: Monte Carlo Dropout confidence assessment quantifies the predicted probability distribution, and the accuracy of identifying high-risk areas exceeds 90%;

[0044] (3) Real-time visualization performance: GPU parallel rendering achieves ≥15fps dynamic heat map refresh, reducing the latency of erosion front location;

[0045] (4) Maintenance cost optimization: AR work order push reduces on-site analysis time and significantly reduces over-maintenance.

[0046] (5) Through the implementation of this invention, tunnel segment operation and maintenance will leapfrog towards digitalization and intelligence, providing a technical benchmark for infrastructure durability management. Attached Figure Description

[0047] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0048] Figure 1 This is a schematic diagram of a dynamic imaging system for chloride ion erosion based on multimodal sensor fusion according to an embodiment of the present invention.

[0049] Figure 2 This is a flowchart of a dynamic imaging method for chloride ion erosion based on multimodal sensor fusion according to an embodiment of the present invention;

[0050] Figure 3 This is a schematic diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0051] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0052] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0053] This embodiment proposes a dynamic imaging system for chloride ion erosion based on multimodal sensor fusion, such as... Figure 1 ,include:

[0054] Embedded sensor array module: used to realize multi-parameter in-situ sensing of Cl⁻ concentration, pH value, corrosion current, strain and temperature, and to provide initial conditions and boundary constraints for physical equations for multi-field coupling modeling module;

[0055] Multi-field coupling modeling module: used to construct a strongly coupled model of Cl⁻ diffusion-steel corrosion-concrete cracking, extract boundary conditions based on crack topology recognition algorithm, and output corrosion expansion force distribution cloud map;

[0056] Dynamic inversion module: Based on the corrosion expansion force distribution cloud map, it generates a continuous monitoring field through spatiotemporal coding, uses bidirectional LSTM to predict the erosion depth in the future preset time period, and quantifies the prediction reliability based on the Monte Carlo Dropout method;

[0057] 3D visualization engine module: used to build a digital twin base, realize dynamic rendering of red-yellow-green erosion heat map, and provide quantitative indicators;

[0058] Mobile inspection interaction module: It is used to overlay erosion heat maps and steel deformation data on smart glasses through SLAM positioning and lightweight rendering engine. It supports voice command to generate structured inspection logs and offline operations, and automatically pushes maintenance work orders containing treatment priorities and material lists based on erosion risk level.

[0059] Specifically, the in-situ monitoring data acquired by the embedded sensor array module provides the initial conditions and boundary constraints for the physical equations of the multi-field coupled modeling module; the physical field states, such as the corrosion expansion force cloud map calculated by the multi-field coupled modeling module, are fused and improved into high-resolution erosion predictions by the dynamic inversion module; the prediction results drive the 3D visualization engine module to generate dynamic heat maps and quantitative indicators; finally, the mobile inspection interaction module uses AR technology to transform the visualization results into on-site executable operation and maintenance instructions. At the same time, prediction errors and on-site verification data form two key feedback links, triggering the recalculation of the physical model and the retraining of the prediction algorithm, respectively, thereby realizing dynamic optimization of the entire process from perception, diagnosis, prediction to decision-making.

[0060] Furthermore, the embedded sensor array module includes:

[0061] Multi-parameter sensing unit: used to deploy MEMS electrochemical sensor arrays according to the spacing of the pipe segment joints, capture Cl⁻ concentration, pH value and corrosion current data, and simultaneously deploy distributed optical fibers to measure strain and temperature data;

[0062] Pressure-resistant encapsulation unit: used to form a protective structure with a silicon carbide ceramic shell and a nano-modified epoxy resin filler layer;

[0063] Array calibration unit: used to dynamically eliminate electromagnetic interference from the tunnel boring machine, and combined with temperature compensation circuit to stabilize the fluctuation range of Cl⁻ concentration measurement within the preset range.

[0064] Specifically, in actual implementation, this embodiment uses an embedded sensor array module to deploy a MEMS sensor array according to the spacing of the pipe segment joints, deploys distributed optical fibers to monitor concrete deformation, and uses a silicon carbide ceramic shell (measured compressive strength 80MPa) and a nano-epoxy resin filling layer to form an IP68 waterproof structure, and transmits data synchronously via an RS485 bus. Through a pulsed low-power sampling strategy, the sensor survival rate in harsh environments is significantly improved, providing a stable data source for dynamic imaging.

[0065] Among them, the multi-parameter sensing unit is used to optimize the spatial arrangement of the microelectrode array, control the MEMS sensor to collect electrochemical parameters at a sampling rate of 1Hz, and synchronously trigger fiber optic strain measurement.

[0066] In actual operation, the multi-parameter sensing unit has a built-in spatial topology self-learning algorithm that automatically plans the sensor deployment density based on the distribution of steel bars in the pipe segment, and uses CRC check to diagnose the data transmission integrity in real time.

[0067] The pressure-resistant encapsulation unit is used to withstand the instantaneous impact load of 80MPa during the assembly of the tunnel boring machine through a silicon carbide ceramic shell, and to block seawater penetration through a nano-modified epoxy resin filling layer (achieving an IP68 waterproof rating), forming a double protection structure.

[0068] The array calibration unit is used to run an environmental noise learning algorithm to dynamically compensate for temperature drift errors, so that the Cl⁻ measurement fluctuation is stabilized within ±5%.

[0069] Temperature compensation circuits are specialized electronic circuit designs whose core function is to automatically identify and offset the negative impact of ambient temperature changes on sensor measurement accuracy. This ensures that the sensor output signal reflects only the true changes in the measured target parameter (such as Cl⁻ concentration), rather than being "contaminated" by temperature fluctuations. In chloride ion corrosion monitoring, the output signal of the core MEMS electrochemical sensor is significantly affected by ambient temperature: parameters such as the rate of electrochemical reactions, ion mobility, and electrode potential are all closely related to temperature. As temperature increases, the current or voltage signal output by the sensor may increase, even if the Cl⁻ concentration in the solution remains unchanged. Without compensation, in environments with large temperature fluctuations within tunnels, the measured Cl⁻ concentration data will contain significant "temperature noise," leading to data distortion and failing to accurately reflect the corrosion status.

[0070] Furthermore, the multi-field coupling modeling module includes:

[0071] Erosion kinetics unit: used to establish a strongly coupled model of the unsteady Cl⁻ diffusion equation and the steel electrochemical corrosion rate equation, and to introduce the phase field theory of concrete crack propagation to quantify the influence of microcracks on the ion diffusion coefficient.

[0072] Boundary condition element: used to automatically extract the real crack network through the microcrack topology recognition algorithm, as the model boundary condition, and dynamically correct the surface ion concentration boundary value by real-time access to the salinity / humidity data of the weather station.

[0073] Real-time solver unit: Used to decompose the concrete mesh model into multiple computing nodes using a GPU parallel finite element computing architecture. When the fiber optic sensor detects a sudden increase in strain, it automatically triggers recalculation and outputs a cloud map of the expansion force distribution of corrosion products.

[0074] Specifically, in the actual execution process, the erosion dynamics unit first constructs a strongly coupled physical field equation of Cl⁻ diffusion-steel corrosion-crack propagation; the boundary condition unit accesses the strain data of the distributed optical fiber in real time, automatically identifies the microcrack topology network and integrates the salinity data of the meteorological station to correct the boundary values; the real-time solution unit automatically triggers GPU parallel recalculation and outputs the corrosion expansion force distribution cloud map to the dynamic inversion module.

[0075] The strongly coupled model of the unsteady Cl⁻ diffusion equation and the steel reinforcement electrochemical corrosion rate equation describes the entire chloride ion corrosion process by coupling three core physical processes. Among them, chloride ion transport is controlled by the diffusion equation ∂C / ∂t=∇·(D(φ)∇C). Its key innovation lies in constructing the effective diffusion coefficient D(φ) as a function of the concrete phase field damage variable φ, thereby quantifying the accelerating effect of crack propagation on ion transport.

[0076] The concrete mesh model is generated based on a digital twin base. By analyzing the BIM model in the design phase, the geometric topology of the pipe segments and the spatial distribution of the steel reinforcement mesh are automatically obtained. An adaptive hexahedral mesh generation algorithm is used to finely discretize the interface between the concrete cover and the steel reinforcement. At the same time, the measured coordinates of the distributed fiber optic sensors are integrated as the constraints of the mesh nodes. Based on the material parameter library, non-uniform mechanical properties are assigned to different regions. Finally, a three-dimensional finite element calculation model containing corrosion interface elements and sensor embedded nodes is formed.

[0077] Furthermore, the dynamic inversion module includes:

[0078] Spatiotemporal coding unit: used to synchronize sensor network data through the BeiDou timing chip and generate a continuous monitoring field of Cl⁻ concentration using the Kriging spatial interpolation algorithm;

[0079] LSTM prediction unit: used to input a 72-hour spatiotemporal sequence into a bidirectional LSTM neural network and output a prediction of the erosion depth for the next 24 hours, with an absolute error of no more than 2cm;

[0080] Confidence assessment unit: used to generate a probability distribution map of the prediction results using the Monte Carlo Dropout method. When the root mean square error of three consecutive predictions is greater than 2cm, it automatically triggers adaptive retraining of weights.

[0081] Specifically, in actual execution, the spatiotemporal coding unit synchronizes the data streams of each sensor through BeiDou time synchronization and uses the Kriging algorithm to generate a centimeter-level resolution Cl⁻ concentration field; the LSTM prediction unit receives the latest 72-hour spatiotemporal sequence every 30 minutes and outputs the predicted erosion depth for the next 24 hours; when the confidence assessment unit detects that the root mean square error of three consecutive predictions is >2cm, it immediately loads the incremental dataset to start 8 minutes of weight retraining, and the updated model is automatically deployed to the prediction pipeline.

[0082] Furthermore, the 3D visualization engine module includes:

[0083] BIM Fusion Unit: Used to analyze the geometric topology of Revit segment models and bind sensor spatial coordinates with steel reinforcement grid coordinates to build a digital twin base;

[0084] Heatmap generation unit: used to achieve dynamic rendering of ≥15fps based on risk level shading engine, marking areas with erosion depth >5cm in red, 2-5cm in yellow, and <2cm in green, and using transparency gradient technology to overlay historical erosion paths;

[0085] Data-driven unit: used to automatically generate cross-sectional erosion depth map of pipe segment, and output 12 quantitative indicators such as erosion volume growth rate and rebar cross-sectional loss rate by combining the rebar rust expansion deformation simulation algorithm.

[0086] Specifically, in actual implementation, the BIM fusion unit receives sensor coordinate data in real time and maps Cl⁻ concentration / crack width to the digital twin of the steel mesh; the heat map generation unit refreshes the red-yellow-green three-color erosion heat map every 0.5 seconds based on the dynamic inversion results, and superimposes historical erosion paths through transparency gradients; the data-driven unit synchronously generates the erosion profile of the segment section, and pushes 12 indicators to the AR inspection terminal in real time in combination with the rust expansion deformation simulation algorithm.

[0087] The geometric topology is obtained by calling the Revit API interface to traverse the set of entity elements in the segment model, extracting the geometric data of all faces, edges, and vertices and their spatial connection relationships, and converting the Boolean operation relationships and parametric attributes of the components into a lightweight topology description framework according to the international industrial basic class standard.

[0088] The core purpose of building a digital twin base is to establish a dynamic mapping relationship between physical sensors and information models. By binding multi-dimensional data streams such as real-time monitored chloride ion concentration and strain changes with the spatial coordinates of the steel mesh, a data carrier with spatiotemporal consistency is provided for the three-dimensional dynamic inversion and visualization of the erosion process. Ultimately, this supports the data fusion and driving needs of upper-level applications such as corrosion expansion force cloud map rendering and automatic maintenance work order generation.

[0089] Furthermore, the mobile inspection interaction module includes:

[0090] AR overlay unit: used to dynamically overlay erosion heat maps and steel rust expansion deformation data in real time on smart glasses through SLAM positioning system and lightweight rendering engine;

[0091] Human-computer interaction unit: used to respond to voice commands to generate structured inspection logs and automatically enable offline operation mode in tunnel network blind spots;

[0092] Operation and maintenance decision unit: used to automatically mark high-risk tunnel segment components based on real-time erosion data and push customized maintenance work orders containing treatment priorities, recommended processes and material lists.

[0093] Specifically, during actual implementation, when maintenance personnel wear smart glasses to scan the pipe segment, the AR overlay unit matches the BIM coordinates in real time to render the internal erosion status; the human-computer interaction unit records the operation process synchronously and stores it in encrypted form; when the erosion depth is greater than 5cm, the maintenance decision unit immediately triggers a high-risk work order and pushes it to the management terminal.

[0094] In this embodiment, the lightweight rendering engine is a graphics rendering architecture optimized for the limited computing power of mobile and wearable devices. It removes unnecessary geometric details of the BIM model through model mesh simplification, texture compression and detail level technology, reduces the number of drawing calls by using instantiated rendering, and dynamically loads visible area data based on view frustum culling and occlusion query algorithms. This enables real-time fusion of SLAM positioning coordinates and erosion heat map vector layers at a frame rate of ≥15fps on smart glasses, ultimately achieving an immersive interactive experience for on-site maintenance personnel without dizziness.

[0095] The working principle of the embodiments of this application will be described in detail below with a specific example.

[0096] First, the embedded sensor array module is activated to verify the working conditions of the MEMS sensor and the distributed optical fiber, and the array is arranged according to the spacing of the pipe segment joints. A pulsed low-power sampling strategy is used to collect Cl⁻ / pH / corrosion current and strain data. The pressure-resistant encapsulation withstands 80MPa impact, and a dynamic calibration algorithm based on environmental noise learning ensures measurement fluctuations are ≤±5%.

[0097] Next, the multi-field coupling modeling module constructs a strongly coupled model of Cl⁻ diffusion-steel corrosion-concrete cracking: the erosion dynamics unit introduces phase field theory to quantify the influence of microcracks; the boundary condition unit accesses meteorological station salinity data in real time to correct boundary values; when the fiber optic monitors a sudden strain, the real-time solution unit triggers GPU parallel recalculation to output the corrosion expansion force cloud map.

[0098] Secondly, the dynamic inversion module synchronizes data through BeiDou timing, and the spatiotemporal coding unit generates a centimeter-resolution Cl⁻ concentration field; the LSTM prediction unit inputs a 72-hour sequence and outputs a 24-hour erosion depth prediction; when the root mean square error of three consecutive predictions is greater than 2cm, the confidence assessment unit starts an 8-minute weight retraining.

[0099] The 3D visualization engine module integrates the BIM model again, and the heat map generation unit refreshes the red-yellow-green three-color erosion heat map every 0.5 seconds (high risk >5cm is marked in red). The transparency gradient technology overlays historical paths; the data-driven unit outputs 12 indicators such as erosion volume growth rate to the AR terminal in real time.

[0100] Finally, the mobile inspection interaction module uses SLAM to locate the superimposed erosion state on the smart glasses and generates maintenance work orders via voice commands; when the erosion depth is >5cm, a high-risk work order containing a materials list is automatically pushed. After a single cycle, the cache is reset to prepare for the next round of monitoring.

[0101] In summary, this embodiment achieves centimeter-level positioning of erosion fronts and real-time optimization of maintenance strategies through multimodal sensor fusion and dynamic imaging technology, replacing manual sampling and analysis, reducing the risk of structural damage, reducing excessive maintenance costs, and standardizing the operation process. Maintenance personnel can quickly execute decisions through AR interaction, significantly improving the accuracy and intelligence level of marine tunnel durability management.

[0102] This embodiment achieves in-situ sensing of multiple parameters (Cl⁻, pH, strain) and 80MPa impact resistance through an embedded sensor array module; constructs a strongly coupled Cl⁻ diffusion-corrosion-cracking model through a multi-field coupling modeling module; outputs 24-hour erosion depth predictions through a dynamic inversion module; renders dynamic heat maps at ≥15fps through a 3D visualization engine module; and pushes AR maintenance work orders through a mobile inspection interaction module. This significantly reduces reliance on manual sampling and improves the accuracy of erosion early warning and maintenance response efficiency. Thus, it solves the problems of traditional methods such as damaging structural integrity, inability to quantify chain reactions, and insufficient visualization, achieving dynamic tracking of the erosion process and closed-loop optimization of operation and maintenance resources.

[0103] like Figure 2 This embodiment also provides a method for dynamic imaging of chloride ion erosion based on multimodal sensor fusion, including:

[0104] Data on Cl⁻ concentration, pH value, corrosion current, strain, and temperature are acquired through an embedded sensor array module, and measurement accuracy is ensured through pressure-resistant packaging and noise calibration.

[0105] A strongly coupled model of Cl⁻ diffusion-steel corrosion-concrete cracking is constructed using a multi-field coupling modeling module, and the boundary conditions are dynamically corrected and a corrosion expansion force cloud map is output.

[0106] Based on the corrosion expansion force cloud map, a continuous monitoring field is generated through the dynamic inversion module to predict the erosion depth within a preset time period in the future, and the model is retrained when the error exceeds the limit.

[0107] The 3D visualization engine module renders dynamic erosion heatmaps and outputs several quantitative indicators.

[0108] The mobile inspection interaction module presents the erosion status in an AR environment and automatically pushes maintenance work orders.

[0109] By optimizing model parameters based on prediction error feedback, a closed-loop optimization of erosion imaging and operation and maintenance decisions is achieved.

[0110] Specifically, Cl⁻ / pH / corrosion current and strain / temperature data are collected through a multi-parameter sensing unit, a pressure-resistant encapsulation unit withstands 80MPa instantaneous impact and seawater penetration, and an array calibration unit dynamically eliminates noise to ensure that Cl⁻ measurement fluctuation is ≤±5%.

[0111] A strongly coupled model of Cl⁻ diffusion-steel corrosion-concrete cracking was constructed using erosion kinetics elements. Boundary condition elements were used to extract the crack network boundary and meteorological data were incorporated to correct the ion concentration.

[0112] A continuous monitoring field is generated by a spatiotemporal coding unit, and an LSTM prediction unit outputs a 24-hour erosion depth prediction based on a 72-hour sequence. The confidence assessment unit triggers retraining within 8 minutes when the root mean square error of three consecutive predictions is greater than 2cm.

[0113] A digital twin base is constructed through BIM fusion units, and a heat map generation unit renders dynamic heat maps at ≥15fps (>5cm red / 2-5cm yellow / <2cm green). The data-driven unit outputs 12 indicators, including the erosion volume growth rate.

[0114] The AR overlay unit displays the erosion status and steel bar deformation on the smart glasses, and the operation and maintenance decision unit pushes work orders containing material lists according to the risk level.

[0115] By dynamically adjusting the weights of the LSTM network based on prediction error feedback, closed-loop optimization of erosion imaging and maintenance decisions is achieved.

[0116] It should be noted that the foregoing explanation of the embodiment of the chloride ion erosion dynamic imaging system based on multimodal sensor fusion also applies to the real-time prediction and feedback method of this embodiment, and will not be repeated here.

[0117] This embodiment achieves in-situ sensing of multiple parameters (Cl⁻, pH, strain) and 80MPa impact resistance through an embedded sensor array unit; constructs a strongly coupled Cl⁻ diffusion-corrosion-cracking model through a multi-field coupling modeling module; outputs 24-hour erosion depth predictions through a dynamic inversion module; renders dynamic heat maps at ≥15fps through a 3D visualization engine module; and pushes AR maintenance work orders through a mobile inspection interaction module. This significantly reduces reliance on manual sampling and improves the accuracy of erosion early warning and maintenance response efficiency. Thus, it solves the problems of traditional methods such as damaging structural integrity, inability to quantify chain reactions, and insufficient visualization, achieving dynamic tracking of the erosion process and closed-loop optimization of operation and maintenance resources.

[0118] Figure 3 An electronic device provided in this embodiment includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of a dynamic imaging method for chloride ion erosion based on multimodal sensor fusion.

[0119] Specifically, the electronic devices also include:

[0120] A communication interface used for communication between the memory and the processor.

[0121] Memory is used to store computer programs that can run on the processor.

[0122] The memory may include high-speed RAM, and may also include non-volatile memory, such as at least one disk drive.

[0123] If the memory, processor, and communication interface are implemented independently, they can be interconnected via a bus to communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0124] Alternatively, in a specific implementation, if the memory, processor, and communication interface are integrated on a single chip, then the memory, processor, and communication interface can communicate with each other through an internal interface.

[0125] The processor may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0126] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of a dynamic imaging method for chloride ion erosion based on multimodal sensor fusion.

[0127] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A dynamic imaging system for chloride ion erosion based on multimodal sensor fusion, characterized in that, include: Embedded sensor array module: used to realize multi-parameter in-situ sensing of Cl⁻ concentration, pH value, corrosion current, strain and temperature, and to provide initial conditions and boundary constraints for physical equations for multi-field coupling modeling module; Multi-field coupling modeling module: used to construct a strongly coupled model of Cl⁻ diffusion-steel corrosion-concrete cracking, extract boundary conditions based on crack topology recognition algorithm, and output corrosion expansion force distribution cloud map; Dynamic inversion module: Based on the corrosion expansion force distribution cloud map, it generates a continuous monitoring field through spatiotemporal coding, uses bidirectional LSTM to predict the erosion depth in the future preset time period, and quantifies the prediction reliability based on the Monte Carlo Dropout method; 3D visualization engine module: used to build a digital twin base, realize dynamic rendering of red-yellow-green erosion heat map, and provide quantitative indicators; Mobile inspection interaction module: It is used to overlay erosion heat maps and steel deformation data on smart glasses through SLAM positioning and lightweight rendering engine. It supports voice command to generate structured inspection logs and offline operations, and automatically pushes maintenance work orders containing treatment priorities and material lists based on erosion risk level.

2. The dynamic imaging system for chloride ion erosion based on multimodal sensor fusion according to claim 1, characterized in that, The embedded sensor array module includes: Multi-parameter sensing unit: used to deploy MEMS electrochemical sensor arrays according to the spacing of the pipe segment joints, capture Cl⁻ concentration, pH value and corrosion current data, and simultaneously deploy distributed optical fibers to measure strain and temperature data; Pressure-resistant encapsulation unit: used to form a protective structure with a silicon carbide ceramic shell and a nano-modified epoxy resin filler layer; Array calibration unit: used to dynamically eliminate electromagnetic interference from the tunnel boring machine, and combined with temperature compensation circuit to stabilize the fluctuation range of Cl⁻ concentration measurement within the preset range.

3. The dynamic imaging system for chloride ion erosion based on multimodal sensor fusion according to claim 1, characterized in that, The multi-field coupling modeling module includes: Erosion kinetics unit: used to establish a strongly coupled model of the unsteady Cl⁻ diffusion equation and the steel electrochemical corrosion rate equation, and to introduce the concrete crack propagation phase field theory to quantify the influence of microcracks on the ion diffusion coefficient; Boundary condition element: used to extract the boundary of the real crack network through the microcrack topology recognition algorithm, and dynamically correct the boundary value of surface ion concentration by accessing the salinity / humidity data of the weather station in real time; Real-time solver unit: Used to decompose the concrete mesh model into multiple computation nodes using a GPU parallel finite element architecture. When the fiber optic sensor detects a sudden increase in strain, it automatically triggers recalculation and outputs a cloud map of the expansion force distribution of corrosion products.

4. The chloride ion erosion dynamic imaging system based on multimodal sensor fusion according to claim 1, characterized in that, The dynamic inversion module includes: Spatiotemporal coding unit: used to synchronize sensor network data and generate a continuous monitoring field of Cl⁻ concentration using the Kriging spatial interpolation algorithm; LSTM prediction unit: It is used to output the erosion depth prediction for a future time period based on the spatiotemporal sequence of a preset time period input by a bidirectional LSTM neural network, and the absolute error of the prediction depth result does not exceed a preset threshold. Confidence assessment unit: used to generate a prediction probability distribution map, and when the root mean square error of three consecutive predictions is greater than the preset threshold, it automatically loads the incremental dataset to start weight adaptive retraining, and the updated model is automatically deployed to the value prediction pipeline.

5. The dynamic imaging system for chloride ion erosion based on multimodal sensor fusion according to claim 1, characterized in that, The 3D visualization engine module includes: BIM Fusion Unit: Used to analyze the geometric topology of Revit segment models, bind sensor spatial coordinates with steel reinforcement grid coordinates, and build a digital twin base; Heatmap generation unit: used to dynamically render red-yellow-green erosion heatmaps based on the risk level shading engine, and to overlay historical erosion paths using transparency gradient technology; Data-driven unit: used to automatically generate cross-sectional erosion depth diagrams of pipe segments and output quantitative indicators by combining the steel reinforcement rust expansion deformation simulation algorithm.

6. The dynamic imaging system for chloride ion erosion based on multimodal sensor fusion according to claim 1, characterized in that, The mobile inspection interaction module includes: AR overlay unit: used to dynamically overlay erosion heatmaps onto smart glasses using a SLAM positioning system and a lightweight rendering engine; Human-computer interaction unit: used to generate structured inspection logs in response to voice commands; Operation and maintenance decision unit: used to push maintenance work orders according to the erosion risk level.

7. A method for dynamic imaging of chloride ion erosion based on multimodal sensor fusion, applied to the dynamic imaging system for chloride ion erosion based on multimodal sensor fusion as described in any one of claims 1-6, characterized in that, include: Data on Cl⁻ concentration, pH value, corrosion current, strain, and temperature are acquired through an embedded sensor array module, and measurement accuracy is ensured through pressure-resistant packaging and noise calibration. A strongly coupled model of Cl⁻ diffusion-steel corrosion-concrete cracking is constructed using a multi-field coupling modeling module, and the boundary conditions are dynamically corrected and a corrosion expansion force cloud map is output. Based on the corrosion expansion force cloud map, a continuous monitoring field is generated through the dynamic inversion module to predict the erosion depth within a preset time period in the future, and the model is retrained when the error exceeds the limit. The 3D visualization engine module renders dynamic erosion heatmaps and outputs several quantitative indicators. The mobile inspection interaction module presents the erosion status in an AR environment and automatically pushes maintenance work orders. By optimizing model parameters based on prediction error feedback, a closed-loop optimization of erosion imaging and operation and maintenance decisions is achieved.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the dynamic imaging method for chloride ion erosion based on multimodal sensor fusion as described in claim 7.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the dynamic imaging method for chloride ion erosion based on multimodal sensor fusion as described in claim 7.

Images