Steel bar corrosion magnetic induction sensor data comprehensive processing platform based on adjustable distance measurement

Abstract

A comprehensive processing platform based on distance measurement adjustable steel bar corrosion magnetic induction sensor data classifies the distance measurement adjustable steel bar corrosion magnetic induction sensors, compares the classified distance measurement adjustable steel bar corrosion magnetic induction sensors with stored data, analyzes the data on the basis, performs risk assessment on the structure, and finally gives out processing suggestions and feasible schemes. The engineering deployment cost is reduced; the automatic data analysis and decision support system reduces manual intervention and improves the detection efficiency; a real-time risk assessment and early warning mechanism effectively prevents structure collapse accidents, and the service life of the building is prolonged.

Description

Technical Field

[0001] The present invention relates to the technical field of building structure health monitoring, and in particular to a steel bar corrosion detection and data intelligent processing system, specifically a comprehensive processing platform that combines an adjustable distance magnetic induction sensor with big data analysis technology. Background Art

[0002] Worldwide, there are numerous cases of concrete damage caused by steel corrosion, particularly in countries that underwent modernization earlier. Due to a lack of awareness of durability issues, the economic losses caused by steel corrosion are enormous. Therefore, research on non-uniform steel corrosion detection technology is an urgent need for social development. Considering that reinforced concrete structures will continue to be widely used in my country's urbanization and infrastructure construction for a considerable period of time, it is crucial to quantitatively characterize and accurately monitor the degree of steel corrosion in reinforced concrete structures, effectively improve the monitoring accuracy of steel corrosion, and then predict the entire life of reinforced concrete structures and formulate reasonable reinforced concrete structure assessment and maintenance plans based on this, which is of great significance to ensuring the safety of concrete structures.

[0003] Evaluation methods for steel corrosion in concrete are generally categorized into destructive testing (DT) and nondestructive testing (NDT). Traditional DT requires appropriately breaking the reinforced concrete specimen and removing the steel. The extent of embedded steel corrosion is then determined by measuring the weight and mass loss of the steel, or by measuring volume loss or cross-sectional area reduction using 3D laser scanning. While accurate, DT also has significant drawbacks. It relies on destroying the concrete structure, leading to premature cessation of service for components and structures. This significantly limits its widespread application. Non-destructive testing methods are a hot topic of research today. Among them, the steel bar corrosion monitoring method based on the magnetic field principle, Chinese patent application number CN110646505A, the application publication date is January 3, 2020, and the name is "External steel bar corrosion non-destructive monitoring sensor and testing method based on electromagnetic field principle" and Chinese patent application number CN112034035A, the application date is December 4, 2020, and the name is "External steel bar corrosion non-destructive monitoring sensor and testing method based on electromagnetic field principle" and Chinese patent application number CN 113720756 A, the application date is August 25, 2021, and the name is "A magnetic induction sensor for steel bar corrosion with adjustable distance". All three patents provide a steel bar corrosion monitoring sensor externally placed on concrete, which is used to monitor the steel bar corrosion of specific concrete, but the sensor's ranging range is fixed, which makes it difficult to adapt to the detection needs of concrete structures of different thicknesses; data needs to be manually processed after collection, the analysis efficiency is low and is easily affected by human errors; there is a lack of real-time risk assessment and early warning mechanism, and it is impossible to respond to engineering safety hazards in a timely manner. In summary, existing steel corrosion monitoring systems mostly rely on sensors with fixed parameters.

[0004] It can be seen that designing a dynamic ranging adjustment to greatly improve the versatility of the sensor, and a data comprehensive processing platform based on the steel corrosion magnetic induction sensor with adjustable ranging is crucial. Summary of the Invention

[0005] In order to overcome the shortcomings of existing sensors, such as fixed ranging range, difficulty in adapting to the detection needs of concrete structures of different thicknesses, manual processing required after data collection, low analysis efficiency and susceptibility to human errors, lack of real-time risk assessment and early warning mechanism, and inability to respond to engineering safety hazards in a timely manner, the present invention provides a comprehensive processing platform for steel corrosion magnetic induction sensor data with adjustable ranging, which reduces engineering deployment costs; an automated data analysis and decision support system reduces manual intervention and improves detection efficiency; and a real-time risk assessment and early warning mechanism effectively prevents structural collapse accidents and extends the life of buildings.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] A comprehensive data processing platform based on magnetic induction sensors for steel corrosion with adjustable range is proposed. The platform classifies the sensors and compares them with the stored data. On this basis, the platform analyzes the data, conducts a risk assessment on the structure, and finally provides treatment suggestions and feasible solutions.

[0008] The integrated processing platform includes:

[0009] The distance-adjustable steel corrosion magnetic induction sensor uses the principle of electromagnetic induction. The transmitting coil (adjustable frequency range 50-500kHz) generates an alternating magnetic field, and the receiving coil captures the magnetic resistance change signal caused by steel corrosion. A nonlinear relationship model between signal strength and detection depth is established:

[0010]

[0011] Where Srce is the received signal strength, α is the attenuation coefficient related to the dielectric constant of concrete, d is the detection depth, and L is the maximum detection range of the sensor;

[0012] The built-in signal processing module dynamically adjusts the transmit power and receive gain through the PID control algorithm to achieve adaptive adjustment of the detection depth (response time ≤ 50ms, adjustment accuracy ±0.1cm).

[0013] Multi-source data classification processing module, including:

[0014] The data classification unit classifies sensor data into two dimensions: time series (real-time data stream) and spatial dimension (sensor grid coordinates);

[0015] The data comparison unit includes horizontal and vertical comparisons. In the horizontal comparison, the difference between the data and the historical baseline data (the reference signal at the time of initial installation) is analyzed to calculate the corrosion increment. In the vertical comparison, the temperature and humidity sensor data are combined to eliminate the interference of environmental variables on the electromagnetic signal (for example, the signal attenuation rate increases by 3% for every 10°C increase in temperature).

[0016] The data cleaning unit uses a sliding window median filter algorithm (window length = 5 sampling points) to remove noise and realizes multi-sensor data fusion through Kalman filtering (fusion error ≤ 5%).

[0017] Intelligent analysis and decision-making modules include:

[0018] The corrosion feature extraction unit extracts time domain features and calculates the root mean square (RMS) and kurtosis coefficient of the signal; extracts frequency domain features: extracts the main frequency offset in the signal spectrum (reflecting the change in the thickness of the corrosion layer); extracts spatial features: generates a corrosion distribution heat map based on the Kriging interpolation algorithm (spatial resolution 0.5m×0.5m);

[0019] The deep learning prediction model uses a bidirectional LSTM network. Its input features include electromagnetic signal strength, temperature, and historical corrosion volume (time series length = 30 days). The output is the corrosion rate forecast for the next 6 months. Model structure parameters: 2 hidden layers, 128 neurons per layer, using the ReLU activation function; Optimizer: Adam (learning rate = 1e-4, batch size = 64); Loss function: Mean Squared Error (MSE), prediction error ≤ 8%;

[0020] The Structural Safety Assessment Unit, in conjunction with the AISC 360 standard, calculates the residual bearing capacity of steel bars:

[0021]

[0022] Among them, f y is the yield strength of steel bars, A s is the effective cross-sectional area (excluding the amount of rust), d is the thickness of the protective layer, and φ is the safety factor (taken as 0.9).

[0023] The integrated processing platform also includes:

[0024] The adaptive maintenance plan generation module generates maintenance recommendations based on the rule engine and case library, combining risk level and structural location.

[0025] Furthermore, the distance adjustment method of the steel bar corrosion magnetic induction sensor is: automatically adjusting the transmission frequency and receiving gain according to the received signal strength to achieve adaptive detection range.

[0026] The beneficial effects of the present invention are mainly manifested in: dynamic ranging adjustment greatly improves the versatility of sensors and reduces engineering deployment costs; automated data analysis and decision support systems reduce manual intervention and improve detection efficiency; real-time risk assessment and early warning mechanisms effectively prevent structural collapse accidents and extend the life of buildings. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 are an integral part of the present invention.

[0028] Figure 2 It is the original traditional magnetic induction sensor.

[0029] Figure 3 It is the data comprehensive processing platform of the present invention.

[0030] Among them, 11 is an external sensor, 12 is a built-in sensor, 13 is a separate sensor, and 14 is a data integrated processing platform. DETAILED DESCRIPTION

[0031] The present invention will be further described below with reference to the accompanying drawings.

[0032] Reference Figures 1 to 3 , a comprehensive data processing platform for steel corrosion magnetic induction sensors with adjustable ranging, classifies the sensor's adjustable ranging magnetic induction intensity monitoring unit and data processing unit, compares them with the stored data, and analyzes the data on this basis, conducts risk assessment of the structure, and finally gives treatment suggestions and feasible solutions.

[0033] The integrated processing platform includes:

[0034] The distance-adjustable steel corrosion magnetic induction sensor uses the principle of electromagnetic induction. The transmitting coil (adjustable frequency range 50-500kHz) generates an alternating magnetic field, and the receiving coil captures the magnetic resistance change signal caused by steel corrosion. A nonlinear relationship model between signal strength and detection depth is established:

[0035]

[0036] Where Srce is the received signal strength, α is the attenuation coefficient related to the dielectric constant of concrete, d is the detection depth, and L is the maximum detection range of the sensor;

[0037] The built-in signal processing module dynamically adjusts the transmit power and receive gain through the PID control algorithm to achieve adaptive adjustment of the detection depth (response time ≤ 50ms, adjustment accuracy ±0.1cm).

[0038] Multi-source data classification processing module, including:

[0039] The data classification unit classifies sensor data into two dimensions: time series (real-time data stream) and spatial dimension (sensor grid coordinates);

[0040] The data comparison unit includes horizontal and vertical comparisons. In the horizontal comparison, the difference between the data and the historical baseline data (the reference signal at the time of initial installation) is analyzed to calculate the corrosion increment. In the vertical comparison, the temperature and humidity sensor data are combined to eliminate the interference of environmental variables on the electromagnetic signal (for example, the signal attenuation rate increases by 3% for every 10°C increase in temperature).

[0041] The data cleaning unit uses a sliding window median filter algorithm (window length = 5 sampling points) to remove noise and realizes multi-sensor data fusion through Kalman filtering (fusion error ≤ 5%).

[0042] Intelligent analysis and decision-making modules include:

[0043] The corrosion feature extraction unit extracts time domain features and calculates the root mean square (RMS) and kurtosis coefficient of the signal; extracts frequency domain features: extracts the main frequency offset in the signal spectrum (reflecting the change in the thickness of the corrosion layer); extracts spatial features: generates a corrosion distribution heat map based on the Kriging interpolation algorithm (spatial resolution 0.5m×0.5m);

[0044] The deep learning prediction model uses a bidirectional LSTM network. Its input features include electromagnetic signal strength, temperature, and historical corrosion volume (time series length = 30 days). The output is the corrosion rate forecast for the next 6 months. Model structure parameters: 2 hidden layers, 128 neurons per layer, using the ReLU activation function; Optimizer: Adam (learning rate = 1e-4, batch size = 64); Loss function: Mean Squared Error (MSE), prediction error ≤ 8%;

[0045] The Structural Safety Assessment Unit, in conjunction with the AISC 360 standard, calculates the residual bearing capacity of steel bars:

[0046]

[0047] Among them, f y is the yield strength of steel bars, A s is the effective cross-sectional area (excluding the amount of corrosion), d is the thickness of the protective layer, and φ is the safety factor (taken as 0.9);

[0048] The risk levels are divided according to the load-bearing capacity reduction rate, refer to Table 1:

[0049] Bearing capacity reduction rate (%) | Risk Level Maintenance Priority ≤10 Low Regular monitoring 10-30 middle Processing within six months >30 high Fix it now

[0050] The adaptive maintenance plan generation module generates maintenance recommendations based on the rule engine and case library, combining risk level and structural location:

[0051]

[0052] Supports multiple options comparison and recommends the best option through cost-benefit analysis (for example, when the repair cost is greater than 30% of the residual value of the structure, demolition and reconstruction is recommended).

[0053] Risk levels are divided according to the rate of decrease in carrying capacity.

[0054] The distance-adjustable steel bar corrosion magnetic induction sensor of this embodiment measures the degree of steel bar corrosion through the principle of electromagnetic induction and has a built-in signal processing module. It adopts a multi-band signal transmission and reception mechanism and a dynamic distance adjustment algorithm to achieve adaptive adjustment of detection accuracy within a range of 0.5-10cm.

[0055] Data comprehensive processing platform architecture:

[0056] Data acquisition layer: Receives corrosion data and environmental parameters (temperature, humidity, etc.) transmitted by sensors in real time, and supports multi-source data fusion;

[0057] Data storage layer: Based on the integrated data storage architecture, it realizes the unified storage of historical data and real-time data;

[0058] Data processing layer: Cleans data to remove abnormal signals and noise interference; extracts key features such as corrosion rate and corrosion area; uses deep learning models to predict the remaining life of steel bars based on LSTM networks;

[0059] Data analysis layer: Combined with structural mechanics models, it evaluates structural safety and generates risk levels;

[0060] Application layer: Visually display the rust distribution cloud map and output maintenance plans (such as coating repair and steel bar replacement suggestions).

[0061] The steel bar corrosion monitoring method implemented by the platform of this embodiment includes the following steps: a. sensors collect corrosion electromagnetic signals and transmit them to the platform; b. data cleaning and feature extraction; c. corrosion rate prediction based on a deep learning model; d. structural safety risk assessment combined with a mechanical model and output of a maintenance plan.

[0062] 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. A comprehensive data processing platform for steel bar corrosion magnetic induction sensors with adjustable ranging, characterized in that: The steel bar corrosion magnetic induction sensors with adjustable distance are classified and compared with the stored data. On this basis, the data are analyzed, the risk assessment of the structure is carried out, and finally treatment suggestions and feasible solutions are given.

2. The data comprehensive processing platform of the steel bar corrosion magnetic induction sensor based on adjustable ranging according to claim 1 is characterized in that: The integrated processing platform includes: The distance-adjustable steel bar corrosion magnetic induction sensor uses the principle of electromagnetic induction. The transmitting coil generates an alternating magnetic field, and the receiving coil captures the magnetic resistance change signal caused by steel bar corrosion. A nonlinear relationship model between signal strength and detection depth is established: Where Srce is the received signal strength, α is the attenuation coefficient related to the dielectric constant of concrete, d is the detection depth, and L is the maximum detection range of the sensor; Built-in signal processing module dynamically adjusts the transmit power and receive gain through PID control algorithm to achieve adaptive adjustment of detection depth; Multi-source data classification processing module, including: Data classification unit, which classifies sensor data into two dimensions: time series and spatial dimensions; The data comparison unit includes horizontal and vertical comparisons. In the horizontal comparison, a difference analysis is performed with the historical baseline data to calculate the corrosion increment. In the vertical comparison, the temperature and humidity sensor data are combined to eliminate the interference of environmental variables on the electromagnetic signal. The data cleaning unit uses a sliding window median filter algorithm to remove noise and implements multi-sensor data fusion through Kalman filtering; Intelligent analysis and decision-making modules include: The corrosion feature extraction unit extracts time domain features, i.e., calculates the root mean square value and kurtosis coefficient of the signal; extracts frequency domain features, i.e., extracts the main frequency offset in the signal spectrum; and extracts spatial features, generating a corrosion distribution heat map based on the Kriging interpolation algorithm. The deep learning prediction model uses a bidirectional LSTM network. The input features include electromagnetic signal strength, temperature, and historical corrosion amount. The output is the predicted corrosion rate for the next 6 months. The Structural Safety Assessment Unit, in conjunction with the AISC 360 standard, calculates the residual bearing capacity of steel bars: Among them, f y is the yield strength of steel bars, A s is the effective cross-sectional area, d is the thickness of the protective layer, and φ is the safety factor.

3. The data comprehensive processing platform of the steel bar corrosion magnetic induction sensor based on adjustable ranging according to claim 2 is characterized in that: In the deep learning prediction model, the model structure parameters are as follows: hidden layer: 2 layers, 128 neurons per layer, using ReLU activation function; optimizer: Adam, learning rate = 1e-4, batch size = 64; loss function: mean square error, prediction error ≤ 8%.

4. The data comprehensive processing platform of steel bar corrosion magnetic induction sensor based on adjustable ranging as claimed in claim 3 is characterized in that: The integrated processing platform also includes: The adaptive maintenance plan generation module generates maintenance recommendations based on the rule engine and case library, combining risk level and structural location.

5. The data comprehensive processing platform for steel bar corrosion magnetic induction sensors based on adjustable distance measurement as described in claims 2 to 4 is characterized in that: The distance adjustment method of the steel bar corrosion magnetic induction sensor is: automatically adjusting the transmission frequency and the receiving gain according to the received signal strength to achieve self-adaptation of the detection range.

Images