Steel bar corrosion cathode protection intelligent regulation and control method and system based on magnetic induction

Abstract

A steel bar corrosion cathodic protection intelligent regulation and control method based on magnetic induction comprises the following steps that 1, data collection and synchronization are conducted, specifically, magnetic field data around a steel bar are collected in real time through magnetic induction sensor arrays evenly distributed in the circumferential direction, and environment temperature, humidity and chloride ion concentration parameters are synchronously obtained; 2, three-dimensional corrosion parameter analysis is carried out; 3, carrying out dynamic current distribution decision making; 4, performing protection current execution and monitoring; and 5, performing prediction optimization: predicting a corrosion trend in future set time by using an LSTM neural network, and adjusting a protection strategy in advance. The invention further provides a steel bar corrosion cathode protection intelligent regulation and control system based on magnetic induction. According to the method, accurate adaptation of non-uniform corrosion is achieved, energy consumption is optimized, and the local damage risk is avoided.

Description

Technical Field

[0001] The present invention belongs to the field of metal corrosion protection and intelligent sensing technology, and specifically relates to an integrated magnetic induction corrosion detection and adaptive current control method and system, which is used for real-time monitoring of non-uniform corrosion of steel bars in reinforced concrete structures and dynamic distribution of cathodic protection current. Background Art

[0002] In the field of durability maintenance of reinforced concrete structures, cathodic protection technology suppresses steel corrosion by applying external current, but its practical application faces significant challenges. Traditional methods mostly use constant potential or constant current modes, relying on fixed parameters to output current, and cannot adapt to the dynamic needs of non-uniform corrosion. For example, local corrosion areas are underprotected due to insufficient current density, while low corrosion areas are overprotected due to excess current, accelerating concrete peeling and hydrogen embrittlement risks. Although existing technologies have attempted to improve through distributed anodes or electrochemical regulation, the lack of real-time corrosion parameter feedback leads to a serious mismatch between current distribution and corrosion morphology. In addition, corrosion monitoring methods (such as half-cell potential method and linear polarization method) are limited to surface potential or average corrosion rate measurement, and cannot analyze three-dimensional corrosion parameters (depth, area, gradient), making it difficult to support precise regulation.

[0003] Furthermore, fluctuations in ambient temperature and humidity and chloride ion penetration significantly affect the corrosion rate, while traditional systems rely on static compensation curves, and the calibration process is based on laboratory conditions, which cannot adapt to complex working conditions. For example, in a high temperature and high humidity environment, the compensation algorithm ignores the temperature and humidity coupling effect, and the current distribution deviation can reach more than 15%. In addition, the discrete hardware design (separation of sensors, controllers, and power modules) leads to complex installation and high failure rates. The multi-sensor network lacks a coordination mechanism, and the problem of data isolation is prominent. Existing technologies are difficult to achieve rapid response to sudden changes in corrosion, and rely on manual inspections (cycles ≥ 3 months). It is difficult to diagnose faults such as anode passivation and cable short circuits in a timely manner, the maintenance cost is high, and the system reliability is insufficient. Summary of the Invention

[0004] To overcome the shortcomings of existing cathodic protection technologies, such as rigid current distribution, low monitoring dimensionality, and high maintenance costs, the present invention provides a magnetic induction-based intelligent control method and system for cathodic protection of steel bar corrosion. This method uses a magnetic induction sensor array to acquire three-dimensional steel bar corrosion parameters (depth, area, and gradient) in real time. A dynamic response model and machine learning algorithm are combined to achieve adaptive distribution of protection current density. A temperature and humidity compensation module is integrated with a multi-sensor collaborative mechanism to accurately correct for environmental interference. An integrated hardware design incorporates high-precision sensing, a multi-channel potentiostat, and an intelligent control unit. This technology overcomes the limitations of traditional methods, which are static control, low-dimensional monitoring, and poor environmental adaptability. It improves current distribution accuracy to ±5%, reduces energy consumption by over 30%, and significantly extends the service life of structures, providing an efficient and intelligent solution for infrastructure corrosion prevention.

[0005] The technical solution adopted by the present invention to solve its technical problem is:

[0006] A magnetic induction-based intelligent control method for steel bar corrosion cathodic protection comprises the following steps:

[0007] Step 1: Data acquisition and synchronization: A circumferentially evenly distributed magnetic induction sensor array collects real-time magnetic field data around the steel bars, and simultaneously obtains ambient temperature, humidity, and chloride ion concentration parameters;

[0008] Step 2: Analyze 3D corrosion parameters: Based on the improved magnetic moment inversion algorithm, combined with the concrete dielectric constant and carbonization depth constraints, calculate the corrosion volume loss rate, area ratio and gradient distribution, generate a 3D corrosion heat map, and output key corrosion indicators: corrosion depth (accuracy ±3%), high corrosion area (δ ≥ 5%) and gradient mutation point

[0009] Step 3: Make dynamic current distribution decisions: Generate regional current density distribution plans based on corrosion parameters and environmental data through dynamic response models. Constant potential mode: When the maximum corrosion rate is less than 2%, maintain the potential at -0.85V±30mV (vs CSE); Constant current mode: When the corrosion rate is ≥2%, prioritize the current distribution to high corrosion areas according to the gradient (0-300mA / m 2 , error ±5%), superimposed temperature and humidity compensation coefficient, real-time correction of current output to suppress environmental interference;

[0010] Step 4: Execute and monitor the protection current: The multi-channel potentiostat outputs a regulated current to the partitioned anode array, monitors the local potential in real time, and triggers overvoltage protection (threshold -1.2V) to prevent hydrogen embrittlement risk.

[0011] Step 5: Perform prediction optimization: Use the LSTM neural network to predict the corrosion trend at a set time in the future and adjust the protection strategy in advance.

[0012] Furthermore, in step 2, the improved magnetic moment inversion algorithm is processed as follows: first, the magnetic field distribution and the corrosion parameters are mapped into a nonlinear equation system; Tikhonov regularization is used to enhance the well-posedness of the inverse problem solution, and the objective function Φ = ||B_obs-B_cal|| is constructed. 2 +λ||Lδ|| 2 , where Bobs is the measured magnetic field, Bcal is the calculated magnetic field, L is the Laplace smoothing operator, and λ is the regularization parameter. The concrete dielectric constant ε is introduced to constrain the magnetic permeability distribution, and the carbonization depth d is used as the boundary condition to limit the corrosion area. The finite element method is used to discretize the three-dimensional space, and the conjugate gradient method is used to iteratively solve the problem, outputting the corrosion volume loss rate δ, area A, and gradient.

[0013] Furthermore, in step five, the process of the LSTM neural network prediction method is as follows: constructing a three-layer network structure containing 128 neurons, the input layer receives 168 hours of normalized time series data, namely, corrosion depth, gradient, and temperature and humidity parameters; using the Adam optimizer and mean square error loss function, the model is iteratively trained based on 30-day sliding window data; the real-time data is divided into 24-hour steps and input into the network, and the corrosion rate prediction value for the next 7 days is output, and the prediction error is corrected by Kalman filtering to be less than 5%; when the predicted value exceeds the corrosion rate threshold, the core control module increases the current density in the high gradient area by 10%-20% in advance, and triggers online incremental learning of LSTM parameters based on real-time monitoring data to achieve dynamic optimization of the model.

[0014] In step 1, the sensor array operates at a sampling frequency of 10-100 Hz to ensure real-time and continuity of data.

[0015] In the fourth step, a high catalytic activity coating (IrO2-Ta2O5) is coated on the surface of the anode array to ensure that the current efficiency is ≥95%.

[0016] A magnetic induction-based steel bar corrosion cathodic protection intelligent control system includes a magnetic induction sensor module, an environmental compensation module, a core control module, and an execution and feedback module.

[0017] The magnetic induction sensor module includes an array of magnetoresistive sensors (the number is ≥12 and is an integer multiple of 4) uniformly arranged along the circumference of the steel bar, and the axial spacing is dynamically adjusted according to the diameter of the steel bar (the spacing is 5 times ± 2 times the diameter of the steel bar);

[0018] The environmental compensation module collects environmental parameters in real time through temperature sensors and humidity sensors, calculates dynamic compensation coefficients, and adjusts the temperature compensation coefficient linearly with the ambient temperature, and the humidity compensation coefficient is corrected with humidity changes. At the same time, a chloride ion concentration threshold is introduced (when it is greater than 1.5 mol / L, the current density is increased by 10%-20%).

[0019] The core control module reconstructs the three-dimensional corrosion parameters of the steel bars, including corrosion depth, area ratio, and gradient distribution, through an improved magnetic moment inversion algorithm combined with the dielectric constant of concrete and carbonization depth constraints. The dynamic response model in the core control module is a corrosion rate-current density dynamic response model, which can generate a regional current density distribution scheme based on real-time corrosion parameters (corrosion depth, gradient) and environmental data (temperature, humidity). The current density distribution adopts a nonlinear function to give priority to compensating high-corrosion areas (when the corrosion rate is ≥2%, the current density is allocated to 0-300mA / m 2 ), and dynamically correct the output through the temperature and humidity compensation coefficient.

[0020] The execution and feedback module includes a multi-channel constant potential instrument and a partitioned anode array.

[0021] Preferably, the sensor has a sensitivity of no more than 0.5 nanotesla and is encapsulated in a corrosion-resistant housing with an IP68 protection grade. The housing material is an epoxy resin composite material that can withstand concrete carbonization, chloride ion corrosion, and extreme environments of -20°C to 80°C.

[0022] The sensor collects magnetic field data in real time at a sampling frequency of 10-100 Hz, and synchronizes with the environmental compensation module to obtain ambient temperature (accuracy ±0.5°C), humidity (range 0-100% RH, error ±2%) and chloride ion concentration (resolution 0.01 mol / L).

[0023] Each sensor node shares data through the cloud platform, uses a federated learning algorithm to optimize the global current distribution strategy, and prioritizes computing resources in key areas.

[0024] Furthermore, the core control module includes a long short-term memory (LSTM) neural network module, which inputs historical corrosion data and environmental parameters, predicts the corrosion development trend in the next 7 days (prediction error is less than 5%), and triggers online update of model parameters.

[0025] In the core control module, the spatial resolution reaches 1 cubic millimeter and the inversion accuracy is controlled within ±3%.

[0026] The current density distribution scheme is as follows: when the maximum corrosion rate is less than 2%, the system maintains a constant potential mode (-0.85V ± 30mV, vs CSE); when the corrosion rate exceeds 2%, it automatically switches to a constant current mode and dynamically adjusts the current density according to the corrosion gradient.

[0027] The multi-channel potentiostat supports 8-16 independent outputs, with current density control accuracy of ±5%, potential control accuracy of ±0.5 millivolts, and a response time of less than 10 milliseconds. Each channel is equipped with an overvoltage protection circuit (trigger threshold -1.2 volts) to prevent the risk of hydrogen embrittlement.

[0028] The partitioned anode array uses a titanium-based mixed metal oxide (MMO) anode strip (width 10-20 cm) with an iridium oxide-tantalum oxide (IrO2-Ta2O5) catalytic layer coated on the surface. The current efficiency is not less than 95% and the design life is more than 20 years.

[0029] In response to the shortcomings of traditional steel corrosion monitoring technology, the present invention proposes an intelligent cathodic protection control method and system based on magnetic induction. The method collects steel bar magnetic field data in real time through a circumferentially distributed magnetic induction sensor array, and combines an improved magnetic moment inversion algorithm to analyze three-dimensional corrosion parameters, breaking through the limitation of traditional potential method and resistance method that can only obtain surface corrosion information; constructs a corrosion rate-current density dynamic response model, integrates LSTM neural network prediction and environmental temperature and humidity compensation mechanism, and realizes adaptive distribution and advance control of protection current; adopts an integrated hardware design, integrates high-precision sensors and multi-channel constant potentiostats to solve the problems of rigid current distribution, poor environmental adaptability and high maintenance cost of traditional systems.

[0030] The beneficial effects of the present invention are primarily manifested in: dynamic control of electrochemical repair parameters, breaking through the traditional static repair model and achieving precise adaptation to non-uniform corrosion; intelligent allocation of repair resources based on corrosion severity and structural criticality, optimizing energy consumption and avoiding the risk of localized damage; adaptive control combined with temperature and humidity compensation, reducing ineffective energy consumption by 30%-40% and extending anode life; and significantly improving system robustness and reducing maintenance costs through corrosion trend prediction and multi-sensor collaboration. This invention is suitable for complex environments such as cross-sea bridges and underground facilities, extending the service life of structures by 5-10 years, and providing an efficient and economical innovative solution for infrastructure corrosion prevention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a flow chart of the intelligent control method for cathodic protection of steel corrosion based on magnetic induction.

[0032] Figure 2 This is a block diagram of the module composition of the intelligent control system for cathodic protection of steel corrosion based on magnetic induction.

[0033] Figure 3 Schematic diagram of data flow coordination of the main modules of the present invention. DETAILED DESCRIPTION

[0034] The present invention will be further described below with reference to the accompanying drawings.

[0035] Reference Figures 1 to 3 , an intelligent control method for cathodic protection of steel corrosion based on magnetic induction, comprising the following steps;

[0036] Step 1: Data Collection and Synchronization

[0037] Magnetic field data acquisition: Through a circumferentially evenly distributed magnetic induction sensor array (number ≥ 12, axial spacing 5D ± 2D, D is the steel bar diameter), the magnetic field data around the steel bar is collected in real time at a frequency of 10-100Hz.

[0038] Environmental parameter synchronization: Synchronously acquire temperature (±0.5°C), humidity (±2% RH) and chloride ion concentration (resolution 0.01 mol / L) data.

[0039] Step 2: 3D corrosion parameter analysis

[0040] Magnetic moment inversion calculation: Based on the improved magnetic moment inversion algorithm, combined with the concrete dielectric constant and carbonization depth constraints, the corrosion volume loss rate δ(x,y,z,t), area ratio A and gradient are inverted Inversion accuracy ≤±3%.

[0041] The improved magnetic moment inversion algorithm is processed as follows: first, the magnetic field distribution and corrosion parameters are mapped into a nonlinear equation system; Tikhonov regularization is used to enhance the well-posedness of the inverse problem solution, and the objective function Φ = ||B_obs-B_cal|| is constructed. 2 +λ||Lδ|| 2 , where Bobs is the measured magnetic field, Bcal is the calculated magnetic field, L is the Laplace smoothing operator, and λ is the regularization parameter. The concrete dielectric constant ε is introduced to constrain the magnetic permeability distribution, and the carbonization depth d is used as the boundary condition to limit the corrosion area. The finite element method is used to discretize the three-dimensional space, and the conjugate gradient method is used to iteratively solve the problem, outputting the corrosion volume loss rate δ, area A, and gradient. During the inversion process, multi-sensor data is fused in real time, and the solution space is optimized in combination with constraint conditions, ultimately achieving the ±3% accuracy requirement.

[0042] Heat map generation: Output 3D corrosion heat map, identify high corrosion areas (δ≥5%) and gradient mutation points

[0043] Step 3: Dynamic Current Distribution Decision

[0044] Model calculation: Generate regional current distribution scheme through the corrosion rate-current density dynamic response model.

[0045] Mode switching can be divided into:

[0046] Constant potential mode (δ_max<2%): holding potential -0.85V±30mV (vs CSE);

[0047] Constant current mode (δ_max ≥ 2%): distribute current density according to the gradient (0-300mA / m 2 , error ±5%).

[0048] Step 4: Protection current execution and optimization

[0049] The current is output according to the distribution scheme through a multi-channel constant potential instrument, the potential is monitored in real time and overvoltage protection (threshold -1.2V) is triggered; the surface of the anode array is coated with a high catalytic activity coating (IrO2-Ta2O5) to ensure that the current efficiency is ≥95%.

[0050] Step 5: Prediction Optimization:

[0051] The LSTM neural network is used to predict the corrosion trend in the next 7 days (inputting 168 hours of historical data, the prediction error is less than 5%), and the model parameters are dynamically adjusted.

[0052] The LSTM neural network prediction method comprises the following steps: constructing a three-layer network structure containing 128 neurons, wherein the input layer receives 168 hours of normalized time series data (corrosion depth, gradient, and temperature and humidity parameters); using the Adam optimizer and the mean square error loss function, iteratively training the model based on 30-day sliding window data; dividing the real-time data into 24-hour steps and inputting it into the network, outputting a predicted value of the corrosion rate for the next seven days, and correcting the prediction error by Kalman filtering to less than 5%; when the predicted value exceeds the corrosion rate threshold, the core control module increases the current density in the high-gradient area by 10%-20% in advance, and triggers online incremental learning of the LSTM parameters based on real-time monitoring data to achieve dynamic optimization of the model.

[0053] An intelligent control system for cathodic protection of steel corrosion based on magnetic induction, including a magnetic induction sensor module, is used for data acquisition and preprocessing. Specific technical details include:

[0054] Sensor array: 12-24 magnetoresistive sensors (sensitivity ≤ 0.5nT), axial spacing 5D ± 2D, encapsulated in an IP68 protective housing;

[0055] Synchronous acquisition: temperature, humidity, and chloride ion concentration data are synchronized in real time, with a sampling frequency of 10-100Hz;

[0056] Data filtering: Sliding average (window = 10) and 3σ principle were used to remove outliers.

[0057] The core control module is responsible for corrosion analysis, model calculation, and strategy optimization. Specific technical details include:

[0058] Algorithm layer:

[0059] Improved magnetic moment inversion algorithm (Tikhonov regularization);

[0060] Dynamic response model (α = 0.2-0.4, β = 0.08-0.15, λ = 3.0-4.5);

[0061] LSTM prediction network (input 168×6 time series data, output 7×1 corrosion trend).

[0062] Computational layer:

[0063] FPGA+ARM dual processor architecture, rust analysis speed ≥1000 points / second, model response time <50ms.

[0064] The environmental compensation module is used to suppress environmental interference and dynamically correct current distribution. The specific technical details are as follows:

[0065] Temperature compensation algorithm: kT = 1 + 0.018 × (T - 20 ° C);

[0066] Humidity compensation algorithm: kh = 1 + 0.025 × (RH - 60%);

[0067] Threshold response algorithm: When Cl- concentration is greater than 1.5 mol / L, enhanced protection is triggered (current +10%-20%);

[0068] The execution and feedback module provides current output, status monitoring, and fault handling. Specific technical details include:

[0069] Potentiostat: 16-channel independent output (±0.5mV accuracy), overvoltage protection threshold -1.2V;

[0070] Anode array: Titanium-based MMO anode strip (IrO2-Ta2O5 coating, life ≥ 20 years).

[0071] The key module coordination mechanism of this embodiment includes:

[0072] Data Flow Collaboration Figure 3 ),as follows:

[0073] Sensing-control: The magnetic induction module transmits the raw data (Bo, T, RH, Cl-) to the core control module;

[0074] Control-compensation: The core control module outputs the basic current distribution scheme, and the environmental compensation module superimposes the correction coefficient;

[0075] Compensation-execution: The corrected current plan is sent to the execution module, and the potentiostat outputs according to the instructions;

[0076] Execution-Feedback: Monitoring data (potential, impedance spectrum) is fed back to the control module in real time, triggering model updates or fault handling.

[0077] Dynamic optimization collaboration is as follows:

[0078] Prediction-driven control: LSTM predicts future corrosion trends, and the core control module adjusts the current density in high-risk areas in advance;

[0079] Environmental Adaptation: The temperature and humidity compensation module corrects the model output in real time, and the linkage execution module enhances protection when the chloride ion exceeds the standard;

[0080] Fault tolerance: When the execution module detects an anomaly, it will feedback to the control module to switch to the backup strategy and push the alarm information to the management platform.

[0081] The resource allocation coordination is as follows;

[0082] Computing resource priority: more computing resources are allocated to high corrosion areas (δ ≥ 5%) to improve analysis and response speed;

[0083] Communication resource optimization: Critical data (such as delta mutations and fault alarms) are transmitted first, and non-critical data are uploaded at a low frequency.

[0084] The embodiments of this specification are merely examples of implementations of the invention and are provided for illustrative purposes only. The scope of protection of the present invention should not be considered limited to the specific embodiments described in these embodiments. The scope of protection of the present invention also extends to equivalent technical means that can be conceived by a person of ordinary skill in the art based on the invention.

Claims

1. An intelligent control method for cathodic protection of steel bar corrosion based on magnetic induction, characterized in that: The method comprises the following steps: Step 1: Data acquisition and synchronization: A circumferentially evenly distributed magnetic induction sensor array collects real-time magnetic field data around the steel bars, and simultaneously obtains ambient temperature, humidity, and chloride ion concentration parameters; Step 2: Analyze 3D corrosion parameters: Based on an improved magnetic moment inversion algorithm, combined with the concrete dielectric constant and carbonization depth constraints, calculate the corrosion volume loss rate, area ratio, and gradient distribution, generate a 3D corrosion heat map, and output key corrosion indicators: corrosion depth, high corrosion areas, and gradient mutation points; Step 3: Make dynamic current distribution decisions: Based on corrosion parameters and environmental data, a dynamic response model is used to generate a regional current density distribution plan. In constant potential mode, when the maximum corrosion rate is less than 2%, the potential is maintained at -0.85V±30mV. In constant current mode, when the corrosion rate is ≥2%, the current is allocated to the high-corrosion area according to the gradient. The temperature and humidity compensation coefficient is added to correct the current output in real time to suppress environmental interference. Step 4: Execute and monitor the protection current: The multi-channel potentiostat outputs a regulated current to the partitioned anode array, monitors the local potential in real time, and triggers overvoltage protection to prevent hydrogen embrittlement risks. Step 5: Perform prediction optimization: Use the LSTM neural network to predict the corrosion trend at a set time in the future and adjust the protection strategy in advance.

2. The intelligent control method for cathodic protection of steel bar corrosion based on magnetic induction according to claim 1, characterized in that: In the step 2, the improved magnetic moment inversion algorithm processing process is as follows: first, the magnetic field distribution and the corrosion parameters are mapped into a nonlinear equation system; Tikhonov regularization is used to enhance the well-posedness of the inverse problem solution, and the objective function Φ = ||B_obs-B_cal||2+λ||Lδ||2 is constructed, where Bob_obs is the measured magnetic field, B_cal is the calculated magnetic field, L is the Laplace smoothing operator, and λ is the regularization parameter; the concrete dielectric constant ε is introduced to constrain the magnetic permeability distribution, and the carbonization depth d is used as a boundary condition to limit the corrosion area; the three-dimensional space is discretized by the finite element method, and the conjugate gradient method is used to iteratively solve the problem, and the corrosion volume loss rate δ, area A and gradient are output.

3. The intelligent control method for cathodic protection of steel bar corrosion based on magnetic induction according to claim 1 or 2, characterized in that: In step 5, the LSTM neural network prediction method comprises the following steps: constructing a three-layer network structure containing 128 neurons, wherein the input layer receives 168 hours of normalized time series data, namely, corrosion depth, gradient, and temperature and humidity parameters; employing an Adam optimizer and a mean square error loss function to iteratively train the model based on 30-day sliding window data; segmenting the real-time data into 24-hour steps and inputting the data into the network, outputting a predicted value of the corrosion rate for the next 7 days, and correcting the prediction error by a Kalman filter to less than 5%; When the predicted value exceeds the corrosion rate threshold, the core control module increases the current density in the high-gradient area by 10%-20% in advance, and triggers online incremental learning of LSTM parameters based on real-time monitoring data to achieve dynamic optimization of the model.

4. The intelligent control method for cathodic protection of steel bar corrosion based on magnetic induction according to claim 1 or 2, characterized in that: In the step 1, the sensor array operates at a sampling frequency of 10-100 Hz.

5. The intelligent control method for steel bar corrosion cathodic protection based on magnetic induction according to claim 1 or 2, characterized in that: In the fourth step, a high catalytic activity coating is coated on the surface of the anode array.

6. A system for implementing the intelligent control method for cathodic protection of steel corrosion based on magnetic induction according to claim 1, characterized in that: The system includes a magnetic induction sensing module, an environmental compensation module, a core control module and an execution and feedback module; The magnetic induction sensor module includes an array of magnetoresistive sensors evenly arranged along the circumference of the steel bar, and the axial spacing is dynamically adjusted according to the diameter of the steel bar; The environmental compensation module collects environmental parameters in real time through temperature sensors and humidity sensors, calculates dynamic compensation coefficients, and adjusts the temperature compensation coefficient linearly with the ambient temperature, and the humidity compensation coefficient is corrected with humidity changes, while introducing a chloride ion concentration threshold. The core control module uses an improved magnetic moment inversion algorithm, combined with the concrete dielectric constant and carbonization depth constraints, to reconstruct the three-dimensional corrosion parameters of the steel bars, including corrosion depth, area ratio, and gradient distribution. The dynamic response model in the core control module is a corrosion rate-current density dynamic response model, which can generate a regional current density distribution scheme based on real-time corrosion parameters and environmental data. The current density distribution adopts a nonlinear function to prioritize compensation for high-corrosion areas, and dynamically corrects the output through the temperature and humidity compensation coefficient. The execution and feedback module includes a multi-channel constant potential instrument and a partitioned anode array.

7. The system according to claim 6, wherein: The sensor has a sensitivity of no more than 0.5 nanotesla and is encapsulated in a corrosion-resistant casing with an IP68 protection rating. The casing is made of epoxy resin composite material, which can withstand concrete carbonization, chloride ion corrosion, and extreme environments ranging from -20°C to 80°C. The sensor collects magnetic field data in real time at a sampling frequency of 10-100 Hz, synchronizes with the environmental compensation module, and obtains ambient temperature, humidity, and chloride ion concentration. Each sensor node shares data through a cloud platform, uses a federated learning algorithm to optimize the global current distribution strategy, and prioritizes the allocation of computing resources to key areas.

8. The system according to claim 4 or 5, characterized in that The core control module includes a long short-term memory (LSTM) neural network module, which inputs historical corrosion data and environmental parameters, predicts the corrosion development trend in the next 7 days, and triggers online updates of model parameters.

9. The system according to claim 4 or 5, characterized in that In the core control module, the spatial resolution reaches 1 cubic millimeter, and the inversion accuracy is controlled within ±3%; the current density distribution scheme is: when the maximum corrosion rate is lower than 2%, the system maintains a constant potential mode; when the corrosion rate exceeds 2%, it automatically switches to a constant current mode and dynamically adjusts the current density according to the corrosion gradient.

10. The system according to claim 4 or 5, characterized in that The multi-channel constant potentiostat supports 8-16 independent outputs, with a current density control accuracy of ±5%, a potential control accuracy of ±0.5 millivolts, a response time of less than 10 milliseconds, and an overvoltage protection circuit configured for each channel; the partitioned anode array uses a titanium-based mixed metal oxide anode strip coated with an iridium oxide-tantalum oxide catalytic layer.

Images