Pulsed eddy current based on-line monitoring system and method for corrosion defects in metallic pipes and equipment

Abstract

The present application belongs to the technical field of oil and gas storage and transportation and petrochemical equipment corrosion detection, and discloses a metal pipeline and equipment corrosion defect online monitoring system and method based on pulse eddy current. The system comprises a hardware and a software system; the hardware comprises a pulse eddy current surface detection sensor unit and an online monitor, the sensor unit is internally integrated with a motor operation track, a stepping motor and a multi-sensor synchronous motion module, and adopts a "walk-stop-sampling-walk" time-sharing work mode; the software comprises data processing and analysis, corrosion early warning and system display modules. The present application supports real-time / non-real-time differentiated monitoring, can match monitoring strategies according to corrosion rates, realizes automatic monitoring from point measurement to surface coverage, intuitively displays corrosion distribution and evolution trend through multi-layer penetration imaging, differential analysis and three-dimensional morphology reconstruction, adopts explosion-proof packaging, realizes high-precision and high-reliability operation, and solves the problems of small monitoring range, high cost and inability of automatic surface monitoring in the prior art.

Description

Technical Field

[0001] This invention relates to the field of corrosion detection technology for oil and gas storage and transportation and petrochemical equipment, and particularly to an online monitoring system and method for corrosion defects in metal pipelines and equipment based on pulsed eddy currents. Background Technology

[0002] As a pillar industry of the national economy, the petrochemical industry's production processes heavily rely on various static metal equipment such as metal pipelines, storage tanks, and reaction vessels. These devices operate under harsh conditions of high temperature, high pressure, and flammable / explosive environments, making their metal walls highly susceptible to various defects such as uniform corrosion, pitting, and stress corrosion. Because corrosion defects are highly concealed, develop slowly, and exhibit highly random evolution patterns, failure to monitor them promptly and implement effective control measures will directly lead to pipeline leaks, equipment failures, and ultimately, major safety accidents such as fires and explosions, causing severe environmental pollution and huge economic losses. Therefore, conducting online corrosion monitoring of metal pipelines and equipment has become a core guarantee for safe production in petrochemical enterprises.

[0003] However, current corrosion monitoring technology for static equipment in the petrochemical industry still has significant shortcomings. Unlike dynamic equipment such as pumps and compressors, which have achieved real-time online monitoring of parameters such as temperature, noise, and vibration, static equipment is constrained by factors such as large monitoring coverage area, slow corrosion change rate, and the strong randomness of defect generation and development. Existing corrosion monitoring technologies struggle to achieve differentiated, low-cost, automated, and fully covered online monitoring.

[0004] Currently, the following technologies are mainly used for corrosion monitoring of static equipment in petrochemical enterprises:

[0005] The first category is ultrasonic point measurement technology. This technology can accurately detect the wall thickness at a single point and is currently a widely used real-time monitoring method. However, this technology has significant limitations: firstly, it can only complete single-point data acquisition, with a small coverage area, and cannot reflect the overall corrosion distribution of the equipment, easily missing local corrosion defects; secondly, for large static equipment, the measuring point needs to be moved manually multiple times to complete the inspection, which is not only inefficient but also cannot achieve automated continuous monitoring, making it difficult to meet the core needs of large-area monitoring of static equipment.

[0006] The second category is invasive monitoring technologies, such as corrosion pads and resistance probes. These technologies require sensors to be directly implanted inside the equipment. Although they can obtain accurate corrosion data, they can damage the integrity of the equipment, affect normal production, and have high maintenance costs in the later stages.

[0007] The third category is non-invasive monitoring technologies, such as ultrasonic guided wave and X-ray detection. Although these technologies do not interfere with the production process, they have inherent drawbacks such as low detection efficiency, difficulty in achieving real-time online monitoring, and susceptibility to interference from the on-site environment.

[0008] Pulsed eddy current testing technology, as a novel non-invasive electromagnetic testing technology, boasts unique advantages such as no need for coupling agents, deep detection depth, and non-contact measurement, and has shown promising application prospects in the detection of wall thinning in pressure equipment. However, current mainstream pulsed eddy current testing equipment is still mainly offline and manually operated, with application modes mostly involving sampling or local scanning. It cannot flexibly match various monitoring strategies, such as real-time continuous monitoring and non-real-time patrol monitoring, according to the different corrosion rates of different parts of the equipment. Therefore, it is difficult to meet the urgent need for continuous, real-time, and comprehensive corrosion monitoring of key areas in petrochemical static equipment. Furthermore, when achieving wide-area monitoring, the equipment deployment and maintenance costs are high.

[0009] In summary, the existing technology lacks a corrosion monitoring system and method that can match the monitoring mode according to the corrosion rate, realize automated real-time online monitoring of severely corroded pipe fittings, and realize non-real-time monitoring of slowly corroded pipe fittings through handheld detection to save costs, while having strong anti-interference ability, large detection range, and the ability to intuitively present the corrosion distribution and evolution trend. Summary of the Invention

[0010] This invention provides a method and system for online monitoring of corrosion defects in metal pipelines and equipment based on pulsed eddy currents. It aims to overcome the shortcomings of existing technologies and solve the following technical problems in corrosion monitoring of static metal equipment in the petrochemical industry: existing ultrasonic point measurement technology has a small monitoring range, cannot achieve area coverage, and easily misses local corrosion defects; invasive monitoring technologies damage equipment integrity and affect production; non-invasive technologies such as ultrasonic guided waves and X-ray detection are difficult to achieve real-time online monitoring and are easily affected by environmental interference; mainstream pulsed eddy current detection equipment is still mainly operated offline manually, which cannot meet the needs of continuous, automated, and full-coverage online monitoring of static equipment. Therefore, this invention provides a system and method for online monitoring of corrosion defects in metal pipelines and equipment based on pulsed eddy currents, which can differentiate between real-time and non-real-time monitoring modes, adapt to the detection needs of various corrosion defect types, achieve automated online monitoring of severely corroded pipe fittings in real-time mode, and save on the cost of multiple devices by using handheld equipment in non-real-time mode. It also achieves large-area full coverage, strong anti-interference ability, and can intuitively present the distribution and evolution trend of corrosion.

[0011] To achieve the above objectives, the present invention adopts the following technical solution:

[0012] In a first aspect, the present invention provides an online monitoring system for corrosion defects in metal pipes and equipment based on pulsed eddy currents, comprising a hardware system and a software system.

[0013] The hardware system consists of a pulsed eddy current surface detection sensor unit and a pulsed eddy current online monitoring instrument. The corrosion defects include, but are not limited to, wall thinning defects, body crack defects, and weld crack defects in metal pipes and equipment.

[0014] The pulsed eddy current surface detection sensor unit is an integrated structure encapsulated in an explosion-proof shell to adapt to the harsh working conditions of high temperature, high pressure, and flammable and explosive environments in petrochemical sites. Internally, the sensor unit integrates a motor track, a high-precision stepper motor, and a multi-sensor synchronous motion module. This module integrates at least two arrayed pulse excitation probes and signal receiving probes. The sensor unit is used for on-site data acquisition. The pulse excitation probes emit pulsed eddy current signals containing multiple frequency components, which can penetrate metal walls and generate an induced magnetic field. The signal receiving probes collect the secondary magnetic field attenuation signal generated when the pulse is turned off. This signal carries information about the corrosion status of the metal wall and can effectively identify various corrosion defects in metal pipes and equipment, such as wall thinning defects, body cracks, and weld cracks. Through its array layout and high-precision stepper motor drive, the multi-sensor synchronous motion module can achieve large-area coverage in a single scan, enabling continuous surface scanning of a certain area to be inspected, effectively solving the problem of missed local corrosion defects in traditional point measurements.

[0015] The motor running track is adjustable. By adjusting the track's specifications and shape, the multi-sensor synchronous motion module can switch between multiple modes, including local surface monitoring, circumferential monitoring, and semi-circumferential monitoring. Specifically, the pulse eddy current surface detection sensor unit can be of various types, including but not limited to planar, arc-shaped, or circumferentially enclosed types, to adapt to the detection locations of storage tanks, equipment, pipelines, or connecting pipes. The specifications and shape of the motor running track are adjusted according to the detection location to allow the multi-sensor synchronous motion module to switch to a matching operating mode. For example, for different detection locations such as tanks, equipment, pipelines, and connecting pipes, the track can be adjusted to adapt the sensor unit to a planar, arc-shaped, or circumferentially enclosed type.

[0016] Furthermore, depending on the type of corrosion defect being monitored, the pulsed eddy current surface detection sensor unit can be of various types, including but not limited to conventional surface detection sensor units, high-precision surface detection sensor units, or high-lift detection sensor units, to address corrosion defects including but not limited to wall thickness reduction, weld cracks, and corrosion under the insulation layer. The multi-sensor synchronous motion module can be adapted to different types of dedicated pulsed eddy current surface detection sensors. The coil arrangement and frequency adaptability of different dedicated surface detection sensors are optimized for the corresponding defect types to meet different detection accuracy requirements. In the multi-sensor synchronous motion module, the number and spacing of the pulse excitation probe and signal receiving probe are adjustable; the higher the detection accuracy requirement, the denser the arrangement.

[0017] The pulsed eddy current online monitoring instrument and the pulsed eddy current surface detection sensor unit are connected via a wired communication. The pulsed eddy current online monitoring instrument includes a pulse signal transmission and signal acquisition unit, a motor operation control unit, a data encapsulation unit, a remote data transmission unit, and a power supply. The pulsed eddy current online monitoring instrument is used for signal transmission and acquisition, power supply, motor control, and encapsulation and transmission of acquired data for the pulsed eddy current surface detection sensor unit.

[0018] To achieve differentiated monitoring strategies, the pulse eddy current online monitoring instrument is divided into a real-time pulse eddy current online monitoring instrument and a non-real-time pulse eddy current online monitoring instrument. Under the drive of the real-time pulse eddy current online monitoring instrument and the non-real-time pulse eddy current online monitoring instrument, the pulse eddy current surface detection sensor unit operates in a time-sharing mode.

[0019] The real-time pulsed eddy current online monitoring instrument is installed in a fixed location on-site. It connects to a matching pulsed eddy current surface detection sensor unit to perform high-frequency, automatic, and continuous monitoring of pipe fittings with severe and rapid corrosion development. In this mode, no on-site operation is required after installation; the system operates automatically according to a preset cycle. The data acquisition cycle of the real-time online monitoring mode can be dynamically configured according to the corrosion rate of the pipe fitting, typically set to once per hour, daily, or weekly, enabling timely tracking of areas with rapid corrosion development.

[0020] The non-real-time pulsed eddy current online monitoring instrument requires no fixed installation and adopts a pre-set programmed working mode. During field deployment, multiple pulsed eddy current surface detection sensor units are connected via branches to a hub controller (this device is only for line aggregation and does not include control functions), achieving centralized aggregation of multiple sensor signals. This non-real-time online monitoring mode is suitable for pipe fittings with slow corrosion development and can be inspected by personnel on a monthly or longer cycle, minimizing equipment deployment and maintenance costs while ensuring effective monitoring. A single handheld monitor can be sequentially connected to multiple pre-installed hub controllers or pulsed eddy current surface detection sensor units to quickly complete the inspection of multiple pipe fittings, thereby achieving wide-area monitoring while effectively saving the cost of fixed installation of multiple devices.

[0021] The motor operation control unit controls the high-precision stepper motor to operate in a time-sharing mode. In petrochemical sites, sensor units are typically installed in confined spaces. When the stepper motor is running, its own electromagnetic field radiates into the surrounding space, causing severe electromagnetic interference to the weak pulsed eddy current secondary magnetic field attenuation signal, resulting in a decrease in signal-to-noise ratio and affecting detection accuracy. To solve this problem, this invention adopts a "walk-stop-collect-walk" cyclic workflow. Specifically, the time-sharing mode is as follows: the high-precision stepper motor drives the multi-sensor synchronous motion module to move a preset number of steps and then stops. While the motor is stopped, pulsed eddy current signals are transmitted and data is collected. After the data collection is completed, the high-precision stepper motor starts moving again, forming a "walk-stop-collect-walk" cyclic workflow. By strictly controlling the signal transmission and collection timing during the motor's stop period, electromagnetic interference generated during stepper motor operation is effectively avoided from the source, ensuring the purity of the collected signal. Through the motion detection and real-time data processing fusion architecture, using a collaborative design of programmable motor start / stop control + wireless data transmission + dynamic analysis algorithm, high-frequency continuous monitoring without downtime is achieved, solving the industry pain point that traditional detection methods require downtime for detection. In addition, through the triple protection mechanism of "time isolation (motor start-stop control) + spatial isolation (sensor layout optimization) + signal filtering", the anti-interference performance is greatly improved compared with the traditional magnetic shielding box solution, and the interference signal strength is reduced to the background noise level.

[0022] The remote data transmission unit includes, but is not limited to, a Bluetooth module, a 4G module, a 5G module, an IoT communication module, and a local storage module. The IoT module adopts an innovative architecture.

[0023] 1. Establish a hierarchical wireless network:

[0024] The sensor layer uses the Zigbee protocol (<100ms latency); the control layer uses a 5G NR-U private network (<10ms latency); this architecture enables the system to maintain a 97.3% data integrity rate even in environments with strong electromagnetic interference (compared to a measured packet loss rate of >35% when using only Bluetooth transmission).

[0025] 2. Dynamic bandwidth allocation algorithm:

[0026] The data processing unit automatically adjusts the transmission rate according to the detection mode, for example: regular monitoring mode: 50kbps; corrosion hotspot tracking mode: 2Mbps;

[0027] The local storage module is used to temporarily store collected data when the network is interrupted and automatically resume transmission after the network is restored, ensuring that the data is not lost.

[0028] The software system includes a data processing and analysis module and a system display module.

[0029] The data processing and analysis module is used to perform availability assessment, multi-dimensional feature extraction, multi-layer penetration imaging, differential analysis, three-dimensional morphology display, corrosion rate and remaining lifetime calculation on the received data, and automatically upload the data analysis results and images to the system display module.

[0030] Specifically, the data processing and analysis module includes:

[0031] The data availability assessment unit is used to determine the availability of the acquired voltage decay curves, automatically identifying and removing abnormal data. Considering that the voltage decay trend of pulsed eddy current signals exhibits a stable and repeatable physical law in specific materials, any significant deviation from this law may be caused by sensor malfunction, strong electromagnetic interference, equipment abnormality, or on-site misoperation. This unit employs an anomaly detection model based on an AI large-scale model framework. This model constructs an available signal database using pre-collected voltage decay curve data under normal operating conditions and trains it to determine the availability of real-time acquired data. If the signal is determined to be normal, subsequent data analysis steps are initiated, and the data set is added to the available signal sample library for continuous model optimization and iteration. If the signal is determined to be abnormal, an alarm is triggered.

[0032] A multi-dimensional feature extraction unit is used to extract features from different dimensions for different defect types. First, the original pulsed eddy current signal is a voltage decay curve over time, and its complete process includes three parts: "air layer - metal wall - air layer". Since the signal response in the air layer is unrelated to the metal corrosion state and is considered interference, it must be identified and removed first. By selecting the number of voltage layers, the unit filters out voltage signals that only reflect the interior of the metal wall, eliminating interference from the air layer. Then, a fitting point selection algorithm is used to select the most stable measurement point from all acquisition points as the fitting point. The voltage value of each layer at all acquisition points is divided by the voltage value of the corresponding layer at the fitting point to obtain the feature value of each point, effectively removing the influence of voltage signal dimensions and eliminating signal deviations caused by environmental factors and equipment fluctuations. Specifically, for wall thinning defects, feature values ​​reflecting wall thickness changes are extracted from the effective signal, and these feature values ​​are converted into wall thickness values ​​using a material-specific calibration curve. For crack defects, a decay feature curve reflecting the degree of crack defects is extracted.

[0033] The multi-layer penetration imaging unit is used to reconstruct corrosion distribution images at different depths from feature data and superimpose them to form a three-dimensional corrosion distribution map. This unit is based on the electromagnetic penetration detection principle and reconstructs the medium distribution images at different depths within the monitored object using a multi-source signal inversion algorithm. A total of k probes are deployed in the monitoring area, each simultaneously acquiring pulsed eddy current signals. After processing by the feature extraction unit, each probe generates a set of feature values ​​reflecting different penetration depths (i.e., different time windows). For the Lth layer, the feature values ​​of the k probes are spatially mapped according to their physical arrangement within the equipment box, forming the feature vector for that layer. The system automatically allocates the actual depth range represented by each layer based on the proportional relationship between the time window value corresponding to each layer and the designed wall thickness of the pipe fitting. From the starting layer to the ending layer, a two-dimensional corrosion distribution image for each layer is generated sequentially. The image is a multi-layer penetration imaging map, with the horizontal axis corresponding to the position of the pulsed eddy current surface detection probe, and the vertical axis or color mapping the magnitude of the feature value. The larger the feature value, the more obvious the warning color in the image; a feature value close to a reference value corresponds to green. By stacking the images of each layer in depth order, a three-dimensional corrosion distribution map of the monitored area can be reconstructed, which can intuitively show the distribution characteristics of defects at different depth layers, the three-dimensional morphology of corrosion pits and their evolution trend.

[0034] The image difference analysis unit is used to construct a difference matrix and generate a corrosion trend map by performing difference operations on a point-by-point and time-by-time-window basis, based on the initial acquired data. Specifically, it uses the original dataset acquired during the system's first commissioning or the initial state of the equipment as the benchmark reference system, defining the response value of the i-th measurement point and the j-th time window in the benchmark data as V. ij The response value of the corresponding measurement point and time window in the currently collected data is C. ij Then the difference S ij=V ij -C ij Difference S ij The physical meaning of is the change in the electromagnetic response of the monitored object at the corresponding measuring point and time window position relative to its initial state. This change is directly related to the degree of defect evolution. This is achieved by constructing a difference matrix S=[S ij ] m×n (Where m is the total number of measurement points and n is the number of time window layers), to achieve a full-domain quantitative characterization of defect changes. The difference matrix is ​​imaged using a multi-layer imaging system. The larger the absolute value of the difference, the more severe the corrosion at that location, corresponding to the yellow to red area in the image; when the difference is close to zero, it indicates that there is no significant change at that location, corresponding to the green area.

[0035] The defect 3D morphology display unit is used to stack images of each layer according to their depth ratio to reconstruct the 3D morphology of the defect. Specifically, it reconstructs the 3D morphology of the defect by stacking images of each layer according to their depth ratio. Specifically, it uses the original corrosion area A of each layer calculated by the defect multi-layer penetration imaging unit. k Where k represents the layer number, and k takes values ​​1, 2, ..., q starting from the deepest layer, where q is the total number of physical depth layers; based on a pre-calibrated signal diffusion model in the laboratory, the original corrosion area A of each layer is calculated. k Inverse mapping and scaling are performed to eliminate signal diffusion effects caused by imaging techniques, thus obtaining the corrected true equivalent corrosion area A for each layer. k' Starting from the deepest layer k=1, perform the following processing on the adjacent k-th and (k-1)-th layers in sequence: obtain the wall thickness value h corresponding to the k-th layer. k Linear interpolation is performed between adjacent layers to make the corrosion area parameter change continuously in the thickness direction, thereby constructing the three-dimensional morphology of the defect; based on the interpolated continuous function, combined with the wall thickness h corresponding to each layer... k By performing integration, the volume V of the adjacent interlayer region can be reconstructed. k Complete the linear interpolation and volume V between all adjacent layers. k By summing the results, we obtain the three-dimensional morphology and total volume V of the entire corrosion defect. The formula for calculating the defect volume is: , where Δh k This represents the thickness increment corresponding to the k-th layer. Using this algorithm, the system can accurately quantify the degree of corrosion, providing crucial data for remaining lifetime assessment.

[0036] The corrosion rate calculation unit is used to calculate the corrosion rate. Specifically, based on historical time-series wall thickness data, it performs linear fitting on all wall thickness maps, and calculates the slope as the corrosion rate. The formula for calculating the corrosion rate is: Where CR is the corrosion rate, n is the number of monitoring periods, and t is the corrosion rate. i Let TH be the time of the i-th monitoring. i This represents the wall thickness value at the corresponding time.

[0037] The remaining service life calculation unit is used to calculate and predict the remaining service life of pipe fittings. Specifically, it analyzes the current state using data collected by the equipment, combines this with historical operating data to predict the future degradation rate, and finally conducts an applicability assessment based on relevant standards such as API 579 / ASME FFS / GB / T30579, thereby providing a remaining service life recommendation with a safety factor. The formula for calculating the remaining service life is: Where RUL represents the remaining useful life, TH min For the current minimum wall thickness, TH allow Minimum wall thickness allowed.

[0038] The graded early warning unit is used to calculate a comprehensive risk score based on a weighted comprehensive assessment of the thinning rate, corrosion allowance, and corrosion rate, and to classify alarm levels according to the score threshold.

[0039] This unit collects and analyzes the following three core indicators in real time:

[0040] Thinning rate N1: Reflects the relative degree of thinning of the pipe fitting wall thickness, and is calculated using the following formula: ;

[0041] Corrosion allowance N2: Reflects the remaining margin between the current wall thickness and the minimum allowable wall thickness. The calculation formula is as follows: ;

[0042] Corrosion rate N3: Reflects the rate of wall thickness reduction per unit time, provided by the corrosion rate calculation unit, in mm / year;

[0043] To achieve a comprehensive evaluation across multiple indicators, this unit employs a weighted summation model to calculate the overall risk score: ;

[0044] Wherein, S is the comprehensive risk score, and the higher the value, the greater the risk; N1, N2, and N3 represent the thinning rate, corrosion allowance, and corrosion rate, respectively; ω1, ω2, and ω3 are the weighting coefficients corresponding to each indicator, which can be dynamically adjusted according to factors such as equipment importance and corrosiveness of process media.

[0045] The system display module is used to visualize corrosion-related information, including but not limited to the analysis images, corrosion status, and remaining life of pipe fittings. It also confirms the warning level based on the calculated thinning rate, corrosion allowance, and corrosion rate, and automatically analyzes and provides maintenance recommendations based on the equipment's historical monitoring results.

[0046] As a further description of the above technical solution, the system includes, but is not limited to, having preset dedicated monitoring calibration curves for different materials such as carbon steel, stainless steel, and alloy steel.

[0047] Secondly, the present invention provides an online monitoring method for corrosion defects in metal pipes and equipment based on pulsed eddy currents, applied to any of the systems described above, comprising the following steps:

[0048] Step S1: System Deployment and Parameter Setting

[0049] Adjust the motor's running track to a suitable monitoring mode based on the geometry of the equipment to be inspected and the required monitoring range. Match the corresponding dedicated pulsed eddy current surface detection sensor unit according to the type of corrosion defect to be detected. Select and deploy the appropriate type of pulsed eddy current online monitoring instrument (real-time / non-real-time) based on the corrosion development rate of the pipe fitting to be inspected. Establish a communication connection between the pulsed eddy current online monitoring instrument and the pulsed eddy current surface detection sensor unit, and preset the acquisition time parameters, acquisition period parameters, step parameters, and signal transmission and reception parameters through the pulsed eddy current online monitoring instrument.

[0050] Step S2: Automatic wake-up and time-sharing data acquisition

[0051] For real-time monitoring, the system automatically wakes up at a preset time. The pulse eddy current online monitoring instrument controls a high-precision stepper motor to drive a multi-sensor synchronous motion module to perform a "walk-stop-collect-walk" cyclic working mode along the motor's running track.

[0052] ① The multi-sensor synchronous motion module stops after moving a preset number of steps;

[0053] ② Control the pulse excitation probe to emit a suitable pulse eddy current signal, and control the signal receiving probe to collect the secondary magnetic field attenuation signal;

[0054] ③ After associating the collected data at the current location with the location information and collection time, temporarily store it;

[0055] ④ Repeat the above steps until all preset points have been collected;

[0056] For non-real-time monitoring, the system adopts a pre-defined, programmed operating mode. The operator moves the handheld pulsed eddy current online monitoring instrument to the pipe fitting where the pulsed eddy current surface detection sensor unit is installed. After connecting the monitoring instrument to the sensor unit, no on-site parameter settings or calibrations are required; the equipment automatically executes the data acquisition process according to the preset acquisition parameters (such as step count, signal transmission frequency, etc.). After completing data acquisition for the current pipe fitting, the operator can disconnect and move to the next pipe fitting with a sensor unit to continue the inspection. This mode, through centralized parameter configuration, enables rapid, cyclical inspection of multiple pipe fittings, avoiding repetitive on-site debugging work and significantly improving inspection efficiency.

[0057] Step S3: Data Encapsulation and Remote Transmission

[0058] The device automatically encrypts and encapsulates the collected data and device status information, and sends it to the data server for storage via a remote data transmission unit.

[0059] Step S4: Data Processing and Multidimensional Analysis

[0060] The collected data is processed and analyzed, including: ① assessing the usability of the collected data and removing invalid data with abnormal voltage decay characteristics; ② extracting feature values ​​reflecting wall thickness changes from the valid data; ③ performing multi-layer imaging processing based on the feature values ​​to generate corrosion distribution images at different depths, and superimposing them to form a three-dimensional corrosion distribution map; ④ performing differential analysis between the current corrosion image and the historical baseline image, constructing a difference matrix based on the initial collected data, and generating a multi-layer imaging map of corrosion change trends; ⑤ performing linear fitting based on the wall thickness data of the historical time series to calculate the corrosion rate; ⑥ calculating the remaining service life of the equipment based on the current minimum wall thickness, the minimum allowable wall thickness, and the corrosion rate; ⑦ calculating a comprehensive risk score based on a weighted comprehensive evaluation of the thinning rate, corrosion allowance, and corrosion rate, and classifying the warning level according to the score threshold.

[0061] Step S5: Visualizing the Results

[0062] The multi-layer penetration imaging map, corrosion change trend map, three-dimensional corrosion distribution map, corrosion rate, remaining life and early warning level information generated in step S4 are visualized through the system display module, and maintenance measures suggestions are automatically given based on historical monitoring results.

[0063] Beneficial effects: The online monitoring system and method for corrosion defects of metal pipes and equipment based on pulsed eddy current, through an integrated pulsed eddy current online monitoring instrument, adjustable track and multi-sensor synchronous motion module, can realize multi-mode monitoring of local surface, circumferential or semi-circular. Compared with the single-point measurement method, it can achieve a larger continuous surface coverage, effectively solves the problem that traditional point measurement is prone to missing local corrosion defects, and requires no manual operation throughout the process, with a high degree of automation.

[0064] The aforementioned online monitoring system and method for corrosion defects in metal pipes and equipment based on pulsed eddy currents supports real-time / non-real-time differentiated monitoring modes and can dynamically match monitoring strategies according to the corrosion rate of pipe fittings: for pipe fittings with severe corrosion and a fast corrosion rate, a fixed deployment real-time online monitoring mode is adopted to achieve timely capture and early warning of defect changes; for pipe fittings with slow corrosion, a non-real-time online monitoring mode using handheld devices is adopted, and one device can complete the inspection of multiple pipe fittings without the need to configure fixed devices separately for such pipe fittings, fundamentally saving the purchase and maintenance costs of multiple idle devices, and taking into account both the timeliness of monitoring and cost control.

[0065] The aforementioned online monitoring system and method for corrosion defects in metal pipes and equipment based on pulsed eddy currents achieves long-term continuous monitoring through fixed deployment, fundamentally avoiding the repetitive work of repeated disassembly and reassembly in traditional testing. Taking metal pipes with insulation layers as an example, after the system of this invention is fixedly deployed, there is no need to erect scaffolding, remove and restore the insulation layer, and continuous monitoring for many years can be achieved with a one-time installation, significantly saving labor and material costs.

[0066] The aforementioned online monitoring system and method for corrosion defects in metal pipes and equipment based on pulsed eddy currents adopts a "walk-stop-collect-walk" time-sharing working mode, transmitting and collecting signals after the motor stops, thus avoiding electromagnetic interference generated by the stepper motor from the source. Dedicated calibration curves are preset for carbon steel and stainless steel materials, and different specialized surface detection sensors are equipped with corresponding material-specific calibration parameters, achieving triple precision matching of probe, material, and defect type. Combined with an AI large-scale anomaly detection model, this ensures signal purity and accurate analysis.

[0067] The aforementioned online monitoring system and method for corrosion defects in metal pipes and equipment based on pulsed eddy currents can generate corrosion distribution maps, corrosion change trend maps, and three-dimensional morphology at different depths through multi-layer imaging, differential analysis, and three-dimensional morphology reconstruction. It employs dedicated analysis algorithms for signals collected by different specialized probes to intuitively display the degree of wall thinning, the extension depth and distribution range of cracks, and the evolution trend of various defects, providing a comprehensive and accurate basis for corrosion assessment.

[0068] The aforementioned online monitoring system and method for corrosion defects in metal pipes and equipment based on pulsed eddy currents has a built-in early warning module that fits the corrosion characteristics of static equipment. It can achieve fully automatic unattended operation. The system automatically wakes up, collects data, transmits data, and analyzes data according to a preset cycle. No manual intervention is required throughout the process, providing a scientific basis for equipment maintenance and significantly improving equipment safety assurance capabilities.

[0069] The online monitoring system and method for corrosion defects of metal pipes and equipment based on pulsed eddy currents adopts a combination of 5G communication module and local storage to realize high-speed remote data transmission and breakpoint resume, ensuring that data is not lost. The data server supports long-term storage and historical traceability.

[0070] The aforementioned online monitoring system and method for corrosion defects in metal pipes and equipment based on pulsed eddy currents, wherein the pulsed eddy current surface detection sensor unit and online monitoring instrument are encapsulated in an explosion-proof shell, adapting to the high temperature, high pressure, and flammable and explosive working conditions of the petrochemical industry; the number and spacing of sensors are adjustable to meet different detection accuracy requirements; installation is convenient, deployment cost is low, and it is easy to promote in the industry;

[0071] The aforementioned online monitoring system and method for corrosion defects in metal pipelines and equipment based on pulsed eddy currents effectively fills the technological gap in automatic surface monitoring of corrosion in petrochemical static equipment. It has advantages such as low cost, wide coverage, high precision, and intelligence. In particular, for equipment with insulation layers, it eliminates the need for repeated disassembly and reassembly of insulation and scaffolding, resulting in significant social and economic benefits. Attached Figure Description

[0072] Figure 1 This is a schematic diagram of the structure of the pulse eddy current online monitoring instrument in an embodiment of the present invention;

[0073] Figure 2 This is a schematic diagram of the installation process of the real-time pulse eddy current online monitoring instrument in an embodiment of the present invention;

[0074] Figure 3 This is a schematic diagram of the installation process of the non-real-time pulse eddy current online monitoring instrument in an embodiment of the present invention.

[0075] Figure 4 This is a flowchart illustrating the overall workflow of the device in this embodiment of the invention.

[0076] Figure 5 This is a flowchart of data availability determination in an embodiment of the present invention;

[0077] Figure 6 This is a graph showing the performance evaluation results of the signal anomaly detection model in this embodiment of the invention;

[0078] Figure 7 This is a schematic diagram of 4D multilayer imaging corrosion in an embodiment of the present invention;

[0079] Figure 8 This is a diagram showing the test results of the C202 bottom to E210 bend of the alkylation unit in a petrochemical plant in an embodiment of the present invention;

[0080] Figure 9 This is a diagram showing the test results of the outlet pipeline elbow of the EC-203A air-cooled tower at the top of a petrochemical plant for removing n-butane in an embodiment of the present invention.

[0081] Figure 10 This is a diagram showing the results of multi-layer imaging corrosion analysis during initial installation in this embodiment of the invention.

[0082] Figure 11 This is a diagram showing the analysis results of multilayer imaging corrosion detected recently in an embodiment of the present invention;

[0083] Figure 12 This is a graph showing the corrosion change results generated by differential analysis in an embodiment of the present invention;

[0084] Figure 13 This is a schematic diagram of a defect stereoscopic imaging in an embodiment of the present invention. Detailed Implementation

[0085] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0086] Example 1: System Hardware Structure

[0087] like Figure 1 As shown, the present invention provides an online monitoring system for corrosion defects in metal pipes and equipment based on pulsed eddy currents, comprising a hardware system and a software system. The hardware system consists of a pulsed eddy current surface detection sensor unit and a pulsed eddy current online monitoring instrument.

[0088] The pulsed eddy current surface detection sensor unit is an integrated structure encapsulated in an explosion-proof shell to adapt to the harsh working conditions of high temperature, high pressure, and flammable and explosive environments in petrochemical sites. Internally, it integrates a motor running track, a high-precision stepper motor, and a multi-sensor synchronous motion module. The multi-sensor synchronous motion module integrates at least two arrayed pulse excitation probes and signal receiving probes. The pulse excitation probe emits pulsed eddy current signals containing multiple frequency components, which can penetrate metal walls and generate an induced magnetic field. The signal receiving probe collects the secondary magnetic field attenuation signal generated when the pulse is turned off. This signal carries information about the corrosion status of the metal wall and can effectively identify various corrosion defects in metal pipes and equipment, such as wall thinning defects, body cracks, and weld cracks. Through its array layout and stepper motor drive, the multi-sensor synchronous motion module can achieve large-area coverage in a single scan, enabling continuous surface scanning of a certain area to be detected, effectively solving the problem of traditional point measurements easily missing local corrosion defects.

[0089] The motor running track is adjustable. By adjusting the track's specifications and shape, the multi-sensor synchronous motion module can switch between multiple modes, including local surface monitoring, circumferential monitoring, and semi-circumferential monitoring, to adapt to the monitoring needs of static equipment (such as pipelines and storage tanks) with different geometries and specifications. For example, for different detection locations such as tanks, equipment, pipelines, and connecting pipes, the track can be adjusted to adapt the sensor units to planar, curved, or circumferential covering types. In the multi-sensor synchronous motion module, the number and spacing of pulse excitation probes and signal receiving probes are adjustable to meet different detection accuracy requirements; the higher the detection accuracy requirement, the denser the deployment.

[0090] The pulsed eddy current online monitoring instrument and the pulsed eddy current surface detection sensor unit are connected via a wired connection. The pulsed eddy current online monitoring instrument includes a pulsed eddy current signal transmission and acquisition unit, a motor operation control unit, a remote data transmission unit, and a power supply, responsible for powering and controlling the sensor unit, as well as encapsulating and transmitting data. The motor operation control unit, as the core control center of the system, is responsible for coordinating the operation of the stepper motor and the transmission and reception of pulse signals.

[0091] The pulse eddy current online monitoring instrument includes a real-time pulse eddy current online monitoring instrument and a non-real-time pulse eddy current online monitoring instrument. The real-time pulse eddy current online monitoring instrument is installed in a fixed location on-site and connects to a matching pulse eddy current surface detection sensor unit to perform high-frequency automatic continuous monitoring of the tested pipe fittings with severe corrosion development and rapid corrosion rates. The non-real-time pulse eddy current online monitoring instrument does not require fixed installation and is carried out by personnel to the pipe fittings where the pulse eddy current surface detection sensor unit is installed for roving monitoring. In this mode, the pulse eddy current surface detection sensor unit deployed on-site also operates in a "walk-stop-collect-walk" time-sharing mode, transmitting and collecting signals when the motor is stopped to ensure signal quality is consistent with the real-time mode.

[0092] As attached Figure 2 The image shows an installation diagram of a real-time pulse eddy current online monitoring instrument.

[0093] In the diagram, the real-time pulse eddy current online monitoring instrument is securely mounted near the pipeline using a fixed bracket. The monitoring instrument is connected via cables to multiple pulse eddy current surface detection sensor units pre-covered on the pipeline. After installation, the entire system can enter an unattended, automated, continuous monitoring state.

[0094] As attached Figure 3 The diagram shown is an operation schematic of a non-real-time pulse eddy current online monitoring instrument.

[0095] In the diagram, multiple pulsed eddy current surface detection sensor units are pre-deployed on-site and installed on each pipe fitting to be monitored. These sensor units are connected to a central control unit via branch cables (this device is only for line aggregation and does not include control functions). The operator moves a handheld, non-real-time pulsed eddy current online monitoring instrument to the central control unit. After connecting the handheld monitor's data cable to the central control unit's interface, the handheld monitor automatically accesses each sensor unit sequentially according to a preset program, driving each sensor unit to independently complete data acquisition in a "walk-stop-collect-walk" time-sharing working mode, completing the detection point by point. After completing data acquisition for the area covered by the current central control unit, the operator can disconnect and move to the next pipe fitting area with a central control unit installed to continue the inspection.

[0096] The motor operation control unit controls the high-precision stepper motor to operate in a time-sharing mode. Specifically, the high-precision stepper motor drives the multi-sensor synchronous motion module to move a preset number of steps and then stops. While the motor is stopped, pulsed eddy current signals are transmitted and data is acquired. After the acquisition is completed, the motor starts moving again, forming a "walk-stop-acquire-walk" cyclic workflow. By strictly controlling the signal acquisition timing during the motor's stop period, electromagnetic interference generated during stepper motor operation is effectively avoided from the source.

[0097] The remote data transmission unit preferably uses a 5G communication module and has a built-in local storage module to temporarily store collected data during network interruptions and automatically resume transmission after the network is restored, ensuring no data loss. The power supply provides stable power to the entire hardware system and is suitable for the power supply environment of petrochemical sites.

[0098] Example 2: Software System Module

[0099] The software system consists of a data processing and analysis module and a system display module. The functions of each module are as follows:

[0100] Data Processing and Analysis Module: This is the core analysis module, responsible for processing the analyzed corrosion signals, extracting feature parameters, and generating corrosion distribution images. This module contains the following sub-units:

[0101] ① Data availability assessment

[0102] This unit employs a deep learning-based anomaly detection model to determine the availability of the acquired secondary magnetic field attenuation signal (voltage-time curve). Considering that the voltage attenuation trend of pulsed eddy current signals exhibits a stable and repeatable physical law in specific materials (such as carbon steel and stainless steel), any significant deviation from this law may be caused by sensor malfunction, strong electromagnetic interference, equipment abnormalities, or on-site misoperation. Figure 5 As shown, the model construction process is as follows: A large amount of voltage decay curve data under normal operating conditions is collected in advance to construct a usable signal database, including historical field monitoring data and usable data collected on-site. Data is added to the database to continuously increase its richness. Based on the usable signal database, the model is trained using an AI large-scale model framework, enabling the model to fully learn the inherent feature distribution and decay patterns of normal curves, ultimately generating a signal anomaly detection model. Figure 6The performance evaluation results of the constructed signal anomaly detection model are shown in the figure, demonstrating that the model can effectively identify abnormal data under various interferences. In actual monitoring, the collected voltage attenuation curves are input into the trained deep learning model. The model calculates the similarity (or reconstruction error) between the features of the collected data and normal data. If it is determined to be a normal signal, the subsequent data analysis steps are initiated, and the data set is added to the available signal sample library for continuous optimization and iteration of the model. If it is determined to be an abnormal signal, an alarm is triggered, reminding staff to promptly confirm and handle the situation on-site.

[0103] ② Multi-dimensional feature extraction unit

[0104] The feature extraction unit is responsible for extracting characteristic parameters that can quantitatively characterize the changes in metal wall thickness from the valid signals that have passed usability assessment. First, the original pulsed eddy current signal is a voltage decay curve over time, and its complete process includes three parts: "air layer - metal wall - air layer". Since the signal response in the air layer is unrelated to the metal corrosion state and is considered interference, it must be identified and removed first. By selecting the number of voltage layers, signals reflecting only the internal metal wall voltage are filtered out, eliminating interference from the air layer. Then, a fitting point selection algorithm is used to select the most stable measurement point (i.e., the point where no corrosion has occurred) from all acquisition points as the fitting point. The voltage value of each layer at all acquisition points is divided by the voltage value of the corresponding layer at the fitting point to obtain the feature value for each point. This operation effectively removes the influence of the voltage signal's dimension and eliminates signal deviations caused by environmental factors and equipment fluctuations, making the feature values ​​more accurately reflect the corrosion status of the metal wall. For wall thinning defects, feature values ​​reflecting wall thickness changes are extracted from effective signals and converted into wall thickness values ​​by combining material-specific calibration curves; for crack defects, attenuation feature curves reflecting the degree of crack defects are extracted.

[0105] To verify the accuracy of feature extraction, two corroded elbows calibrated with an ultrasonic thickness gauge were selected for inspection. The inspection results are as follows: Figure 8 and Figure 9 As shown.

[0106] Figure 8 The test results are for the C202 bottom to E210 bend of the alkylation unit in a petrochemical plant. The original wall thickness at this location was 5.6 mm. The test results from the system of this invention showed that the wall thickness was reduced by about 10%, with a reduction of 0.56 mm.

[0107] Figure 9 The test results are for the elbow of the EC-203A outlet pipeline of the top air-cooled n-butane removal tower in a petrochemical plant. The original wall thickness at this location was 7.0 mm, and the test revealed a wall thickness reduction of approximately 10%, or 0.7 mm.

[0108] The corrosion at both locations mentioned above is in its early stages. Traditional point measurement methods are prone to missing detection due to sparse point distribution. However, this invention successfully captures minute corrosion signals through array-type surface scanning and differential analysis algorithms, verifying the system's high-sensitivity detection capability for 10% level thinning.

[0109] ③ Defect multilayer penetration imaging

[0110] This unit, based on the principle of electromagnetic penetration detection, reconstructs the medium distribution images at different depths within the monitored object using a multi-source signal inversion algorithm. Assume a total of k probes are deployed in the monitoring area, each simultaneously acquiring pulsed eddy current signals. After processing by the feature extraction unit, each probe generates a set of feature value sequences reflecting different penetration depths (i.e., different time windows). For the Lth layer (corresponding to a specific time window), the feature values ​​of the k probes are spatially mapped according to their physical arrangement within the device housing (e.g., from top to bottom: probe 1, probe 2, ..., probe k), forming the feature vector for that layer. , where f il This represents the feature value of the i-th probe in layer L. Based on the detection requirements of the image analysis, the starting layer L of the image to be analyzed is set. start With termination layer L end The system automatically assigns the actual depth range represented by each layer based on the proportional relationship between the time window value corresponding to each layer and the designed wall thickness of the pipe fitting. Let the designed wall thickness of the pipe fitting be D, and the total number of imaging layers be N, then the depth range corresponding to the Lth layer is... This ensures that the imaging results directly correspond to the physical thickness. Two-dimensional corrosion distribution images are generated sequentially for each layer, from the starting layer to the ending layer. These images are multi-layer penetration images, with the horizontal axis corresponding to the probe's arrangement position and the vertical axis (or color mapping) representing the probe's characteristic value. Larger characteristic values ​​(i.e., more severe wall thinning) correspond to more prominent warning colors (e.g., yellow to red) in the image; characteristic values ​​close to a baseline value correspond to green. By stacking the images in depth order, a three-dimensional corrosion distribution map of the monitored area can be reconstructed, visually displaying the distribution characteristics of defects at different depths, the three-dimensional morphology of corrosion pits, and their evolution trends. Figure 7 As shown, the figure is divided into multiple images, each representing the corrosion status of the pipe at different depths. Green indicates areas where no corrosion has occurred, while yellow indicates areas where corrosion has occurred. This visually displays information such as the corrosion distribution, degree of corrosion, and depth of corrosion of the pipe.

[0111] ④ Image difference analysis unit

[0112] This unit performs pixel-level differential calculations on imaging data from different monitoring periods to quantitatively analyze the dynamic evolution of corrosion defects, providing a basis for corrosion trend prediction and early warning. The specific implementation steps are as follows:

[0113] Baseline Data System Construction: The original dataset collected during the system's initial commissioning or the equipment's initial state serves as the baseline reference system. This dataset covers the full-time window sequence response signals of all preset measurement points of the monitored object, reflecting the intrinsic electromagnetic response characteristics when defects have not occurred or are in their initial state.

[0114] Differential quantization model establishment: In subsequent monitoring cycles, the current raw dataset is collected and differential calculations are performed on the benchmark data point by point and time window by time window. The response value of the i-th measuring point and j-th time window in the benchmark data is defined as V. ij The response value of the corresponding measurement point and time window in the currently collected data is C. ij Then the difference S ij =V ij -C ij Difference S ij The physical meaning of is the change in the electromagnetic response of the monitored object at the corresponding measuring point and time window position relative to its initial state. This change is directly related to the degree of defect evolution. This is achieved by constructing a difference matrix S=[S ij ] m×n Where m is the total number of measurement points and n is the number of time window layers, the whole-domain quantitative characterization of defect changes is realized;

[0115] Defect Comparison Visualization: The difference matrix S is imaged using a multi-layer imaging system. During imaging, the difference data, as a key quantitative indicator of defect characteristics, is mapped to pixel values ​​of different colors, forming an intuitive corrosion distribution image. The larger the absolute value of the difference, the more significant the change in the medium properties at that location relative to the baseline state, i.e., the more severe the corrosion, corresponding to the yellow to red area in the image; when the difference is close to zero, it indicates no significant change at that location, corresponding to the green area. By overlaying multiple images, a three-dimensional corrosion distribution map can be generated, intuitively displaying the depth, shape, and evolution trend of the defects. Figures 10 to 12 As shown, Figure 10 This is a multi-layered imaging of corrosion during initial installation. Figure 11 This is a recent corrosion assessment. Figure 12 The corrosion change results generated by the differential analysis clearly show that, compared to the initial installation, recent inspections have shown significant further corrosion.

[0116] ⑤ Reconstruction of three-dimensional shape

[0117] This unit utilizes multi-layer penetration imaging technology, stacking multiple imaging layers according to depth ratios to form a three-dimensional representation of the defect morphology, intuitively presenting the depth, shape, and distribution of corrosion pits. The specific reconstruction algorithm is as follows:

[0118] First, the original corrosion area A of each layer is calculated based on the defect multilayer penetration imaging unit. k, where k represents the layer number, and k takes values ​​1, 2, ..., q starting from the deepest layer, where q is the total number of physical depth layers;

[0119] Secondly, based on the pre-calibrated signal diffusion model in the laboratory, the original corrosion area A of each layer was analyzed. k Inverse mapping and scaling are performed to eliminate signal diffusion effects caused by imaging techniques, thus obtaining the corrected true equivalent corrosion area A for each layer. k' ;

[0120] Then, starting from the deepest layer k=1, the following processing is performed on the adjacent k-th and (k-1)-th layers in sequence: obtain the wall thickness value h corresponding to the k-th layer. k Linear interpolation is performed between adjacent layers to make the corrosion area parameter change continuously in the thickness direction, thereby constructing the three-dimensional morphology of the defect;

[0121] Finally, based on the interpolated continuous function, combined with the wall thickness h corresponding to each layer... k By performing integration, the volume V of the adjacent interlayer region can be reconstructed. k Complete the linear interpolation and volume V between all adjacent layers. k By summing the results, we obtain the three-dimensional morphology and total volume V of the entire corrosion defect. The formula for calculating the defect volume is: , where Δh k This represents the thickness increment corresponding to the k-th layer;

[0122] like Figure 13 As shown, this is a 3D image of corrosion constructed based on the effective penetration layer of the detection signal after a 15% thinning in a laboratory corrosion experiment. The defect is a cone-shaped corrosion with a width of 5mm, a length of 7mm, and a height of 2mm. The 3D defect display shows a good reproduction of the true morphology of the defect.

[0123] The corrosion rate calculation unit is used to calculate the corrosion rate. Specifically, based on historical time-series wall thickness data, it performs linear fitting on all wall thickness maps, and calculates the slope as the corrosion rate. The formula for calculating the corrosion rate is: Where: CR is the corrosion rate, n is the number of monitoring periods, and t i Let TH be the time of the i-th monitoring. i This represents the wall thickness value at the corresponding time.

[0124] The remaining service life calculation unit is used to calculate and predict the remaining service life of pipe fittings. Specifically, it analyzes the current state using data collected by the equipment, combines this with historical operating data to predict the future degradation rate, and finally conducts an applicability assessment based on relevant standards to provide a remaining service life recommendation with a safety factor. The formula for calculating the remaining service life is: Where RUL represents the remaining useful life, TH minFor the current minimum wall thickness, TH allow Minimum wall thickness allowed.

[0125] The graded early warning unit is used to calculate a comprehensive risk score based on a weighted comprehensive assessment of the thinning rate, corrosion allowance, and corrosion rate, and to classify alarm levels according to the score threshold.

[0126] This unit collects and analyzes the following three core indicators in real time:

[0127] Thinning rate N1: Reflects the relative degree of thinning of the pipe fitting wall thickness, and is calculated using the following formula: ;

[0128] Corrosion allowance N2: Reflects the remaining margin between the current wall thickness and the minimum allowable wall thickness. The calculation formula is as follows: ;

[0129] Corrosion rate N3: Reflects the rate of wall thickness reduction per unit time, provided by the corrosion rate calculation unit, in mm / year;

[0130] To achieve a comprehensive evaluation across multiple indicators, this unit employs a weighted summation model to calculate the overall risk score: ;

[0131] Wherein, S is the comprehensive risk score, and the higher the value, the greater the risk; N1, N2, and N3 represent the thinning rate, corrosion allowance, and corrosion rate, respectively; ω1, ω2, and ω3 are the weighting coefficients corresponding to each indicator, which can be dynamically adjusted according to factors such as equipment importance and corrosiveness of process media.

[0132] The system display module is used to intuitively display the corrosion information such as images, corrosion status, and remaining life of the analyzed pipe fittings. It also confirms the warning level based on the calculated thinning rate, corrosion allowance, and corrosion rate, and automatically analyzes and provides maintenance and repair suggestions based on the equipment's historical monitoring results.

[0133] Example 3: System workflow, such as Figure 4 As shown

[0134] Step S1: System Deployment and Parameter Setting

[0135] Based on the corrosion development rate of the pipe fittings to be inspected, select and deploy the appropriate type of pulsed eddy current online monitoring instrument (real-time / non-real-time). Match the multi-sensor synchronous motion module with a dedicated pulsed eddy current surface detection sensor unit corresponding to the corrosion defect type (such as wall thinning, cracks). Adjust the specifications and shape of the motor running track according to the geometry of the equipment under inspection and the monitoring range requirements, enabling the sensor unit to switch between multiple modes such as local surface monitoring and circumferential monitoring. Establish a communication connection between the pulsed eddy current online monitoring instrument and the sensor unit, and preset the acquisition time parameters, acquisition period parameters, step parameters, and signal transmission and reception parameters through the online monitoring instrument.

[0136] Step S2: Automatic wake-up and time-sharing data acquisition

[0137] In real-time online monitoring mode, the system automatically wakes up at a preset time and controls the high-precision stepper motor to drive the multi-sensor synchronous motion module to perform a "walk-stop-collect-walk" cyclic working mode along the motor running track: ① stop after moving a preset number of steps; ② control the probe to emit and collect signals; ③ temporarily store the collected data of the current point along with the location and time information; ④ repeat until the collection of all preset points is completed.

[0138] For non-real-time monitoring, the system adopts a pre-defined programmed operating mode. The operator moves the handheld pulsed eddy current online monitoring instrument to the pipe fitting where the pulsed eddy current surface detection sensor unit is installed. After connecting the monitoring instrument to the sensor unit (or hub controller), no on-site parameter settings or calibration are required. The equipment automatically drives the sensor unit to execute a "walk-stop-collect-walk" time-sharing operating mode according to the preset acquisition parameters, sequentially completing data acquisition at each point. After completing data acquisition for the current pipe fitting or the area covered by the current hub controller, the operator can disconnect and move to the next location to be monitored to continue operation.

[0139] Step S3: Data Encapsulation and Remote Transmission

[0140] The device automatically encrypts and encapsulates the collected data and device status information, and sends it to the data server for storage via a remote data transmission unit. If the network is interrupted, the local storage module temporarily stores the data, and the transmission will automatically resume once the network is restored.

[0141] Step S4: Data Processing and Multidimensional Analysis

[0142] The server processes and analyzes the collected data, including: ① using an AI model to determine data availability and remove abnormal data; ② extracting feature values ​​reflecting wall thickness changes from valid data; ③ performing multi-layer penetration imaging based on the feature values ​​to generate corrosion distribution images at different depths, and superimposing them to form a three-dimensional corrosion distribution map; ④ performing differential analysis between the current image and historical baseline images to construct a difference matrix and generate a corrosion trend map; ⑤ performing linear fitting based on historical time-series wall thickness data to calculate the corrosion rate; ⑥ calculating the remaining service life of the equipment based on the current minimum wall thickness, the minimum allowable wall thickness, and the corrosion rate; and ⑦ calculating a comprehensive risk score based on a weighted comprehensive evaluation of the thinning rate, corrosion allowance, and corrosion rate, and classifying warning levels according to the score threshold.

[0143] Step S5: Visualizing the Results

[0144] The corrosion early warning module incorporates a comprehensive corrosion early warning mechanism tailored to the corrosion characteristics of static equipment in the petrochemical industry. Based on relevant standards in the petrochemical industry, equipment usage requirements, and the slow and random nature of corrosion changes in static equipment, it performs real-time and intelligent judgment on the analyzed corrosion data.

[0145] After data processing and analysis are completed, the results are uploaded to the software system's visualization interface. This platform supports practical functions such as quick data querying and historical data comparison, facilitating equipment maintenance and data traceability for enterprise technicians, while also allowing enterprise managers to view the equipment's corrosion status in real time.

[0146] Example 4: Material-Specific Calibration and Multi-Mode Monitoring

[0147] Due to the differences in magnetic permeability and electrical conductivity between carbon steel and stainless steel, their responses to pulsed eddy current signals differ. This invention's system has pre-set dedicated monitoring calibration curves for carbon steel and stainless steel respectively. When the feature extraction unit converts feature values ​​into wall thickness values, it automatically calls the corresponding calibration curve based on the material of the device under test, thereby eliminating the impact of material differences on detection accuracy.

[0148] By adjusting the specifications and shape of the motor's running track, the multi-sensor synchronous motion module can achieve scanning along different trajectories. For pipeline equipment, it can be set to a circumferential monitoring mode, allowing the sensors to move along the circumference of the pipeline; for planar structures such as tank walls, it can be set to a local surface monitoring mode, achieving two-dimensional grid scanning; for semi-enclosed structures, it can be set to semi-circular monitoring. This flexibility allows the invention to adapt to the geometry and monitoring needs of various static equipment.

[0149] Example 5: Remote Data Transmission and Resume of Breakpoint

[0150] The remote data transmission unit employs an industrial-grade 5G communication module, enabling high-speed remote transmission of monitoring data with low latency and high stability, unaffected by the complex environment of petrochemical sites. A built-in local storage module temporarily stores collected data during network interruptions and automatically resumes transmission upon network restoration, ensuring no data loss. The data server provides long-term storage and categorized archiving of monitoring data, supporting rapid data retrieval, historical comparison, and corrosion trend prediction, providing comprehensive and reliable data support for the entire lifecycle management of the equipment.

[0151] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An online monitoring system for corrosion defects in metal pipes and equipment based on pulsed eddy currents, characterized in that, This includes both hardware and software systems; The hardware system includes a pulsed eddy current surface detection sensor unit and a pulsed eddy current online monitoring instrument; The pulse eddy surface detection sensor unit is an integrated structure and is encapsulated in an explosion-proof shell. The pulse eddy surface detection sensor unit integrates a motor running track, a high-precision stepper motor and a multi-sensor synchronous motion module. The multi-sensor synchronous motion module integrates at least two pulse excitation probes and signal receiving probes arranged in an array. The pulse eddy surface detection sensor unit is used for on-site data acquisition. The pulsed eddy current online monitoring instrument includes a pulse signal transmission and signal acquisition unit, a motor operation control unit, a remote data transmission unit, and a power supply. The pulsed eddy current online monitoring instrument is used to provide signal transmission, acquisition, power supply, and control functions for the pulsed eddy current surface detection sensor unit, and to encapsulate and transmit the acquired data. The pulsed eddy current online monitoring instrument and the pulsed eddy current surface detection sensor unit are connected to each other via a wired connection. The software system includes a data processing and analysis module and a system display module; The data processing and analysis module is used to perform availability assessment, multi-dimensional feature extraction, multi-layer penetration imaging, differential analysis, three-dimensional morphology display, corrosion rate, remaining lifetime calculation, and early warning level confirmation on the received data, and automatically upload the data analysis results and images to the system display module. The system display module is used to visualize corrosion-related information, including but not limited to the analysis images, corrosion status, and remaining life of pipe fittings. It also automatically analyzes and provides maintenance recommendations based on the warning level and the historical monitoring results of the equipment. The types of pulsed eddy current surface detection sensor units include, but are not limited to, conventional surface detection sensor units, high-precision surface detection sensor units, or high-lift detection sensor units, respectively targeting corrosion defects including, but not limited to, wall thickness reduction, weld cracks in the body, and corrosion of pipelines under the insulation layer; the multi-sensor synchronous motion module is adapted to different types of dedicated pulsed eddy current surface detection sensors. The pulse eddy current online monitoring instrument includes a real-time pulse eddy current online monitoring instrument and a non-real-time pulse eddy current online monitoring instrument. Driven by both the real-time and non-real-time pulse eddy current online monitoring instruments, the pulse eddy current surface detection sensor unit operates in a time-sharing mode. The real-time pulse eddy current online monitoring instrument is installed in a fixed location on-site and connects to the matching pulse eddy current surface detection sensor unit group to perform high-frequency automatic continuous monitoring of the tested pipe fittings with severe and rapid corrosion development. The non-real-time pulse eddy current online monitoring instrument does not require fixed installation; it is handheld by personnel to the pipe fitting where the pulse eddy current surface detection sensor unit is installed. After connection, it drives the pulse eddy current surface detection sensor unit to automatically complete data acquisition according to a preset time-sharing mode, achieving roving monitoring.

2. The online monitoring system for corrosion defects of metal pipes and equipment based on pulsed eddy currents according to claim 1, characterized in that, The pulse eddy current surface detection sensor unit includes, but is not limited to, planar, arc-shaped, or circumferentially enclosed types, to adapt to the detection locations of storage tanks, equipment, pipelines, or connecting pipes, respectively; the specifications and shape of the motor running track are adjusted according to the detection location so that the multi-sensor synchronous motion module switches to the matching working mode.

3. The online monitoring system for corrosion defects of metal pipes and equipment based on pulsed eddy currents according to claim 1, characterized in that, The motor operation control unit controls the high-precision stepper motor to operate in a time-sharing mode. Specifically, the high-precision stepper motor drives the multi-sensor synchronous motion module to move a preset number of steps and then stops. While the motor is stopped, pulse eddy current signals are transmitted and data is collected. After the data collection is completed, the high-precision stepper motor starts moving again, forming a "walk-stop-collect-walk" cyclic workflow.

4. The online monitoring system for corrosion defects of metal pipes and equipment based on pulsed eddy currents according to claim 1, characterized in that, In the multi-sensor synchronous motion module, the number and spacing of the pulse excitation probe and the signal receiving probe are adjustable to adapt to different detection accuracy requirements.

5. The online monitoring system for corrosion defects of metal pipes and equipment based on pulsed eddy currents according to claim 1, characterized in that, The data processing and analysis module includes: The data availability judgment unit is used to judge the availability of the collected voltage decay curves and automatically identify and remove abnormal data. It adopts an anomaly detection model based on the AI ​​large model framework. The anomaly detection model is used to construct an available signal database by collecting voltage decay curve data under normal operating conditions in advance and training it to judge the availability of real-time collected data. The multi-dimensional feature extraction unit is used to extract features from different dimensions for different defect types. Specifically, for wall thinning defects, feature values ​​reflecting wall thickness changes are extracted from the effective signal and converted into wall thickness values ​​by combining the feature values ​​with the material-specific calibration curve. For crack defects, attenuation feature curves that reflect the degree of crack defects are extracted. The defect multi-layer penetration imaging unit is used to reconstruct the corrosion distribution images of different depth layers from the feature data, and superimpose them to form a three-dimensional corrosion distribution map; the image is a multi-layer penetration imaging map, with the horizontal axis corresponding to the position of the pulse eddy current surface detection probe, and the vertical axis or color mapping feature value magnitude; The image difference analysis unit is used to construct a difference matrix and generate a corrosion trend map by performing difference operations point-by-point and time-by-time window operations on the initial acquired data as a reference. Specifically, it uses the raw data acquired by the system for the first time as a reference and defines the response value of the i-th measurement point and the j-th time window in the reference data as V. ij The response value of the corresponding measurement point and time window in the currently collected data is C. ij Then the difference S ij =V ij -C ij Its physical meaning is the change in electromagnetic response of the monitored object at the corresponding measuring point and time window position relative to the initial state. This change is directly related to the degree of defect evolution. By constructing the difference matrix S=[S ij ] m×n Where m is the total number of measurement points and n is the number of time window layers, a full-domain quantitative characterization of defect changes is achieved; The defect 3D morphology display unit is used to stack images of each layer according to their depth ratio to reconstruct the 3D morphology of the defect; specifically, it uses the original corrosion area A of each layer calculated by the defect multi-layer penetration imaging unit. k Where k represents the layer number, and k takes values ​​1, 2, ..., q starting from the deepest layer, where q is the total number of physical depth layers; based on a pre-calibrated signal diffusion model in the laboratory, the original corrosion area A of each layer is calculated. k Inverse mapping and scaling are performed to eliminate signal diffusion effects caused by imaging techniques, thus obtaining the corrected true equivalent corrosion area A for each layer. k' Starting from the deepest layer k=1, perform the following processing on the adjacent k-th and (k-1)-th layers in sequence: obtain the wall thickness value h corresponding to the k-th layer. k Linear interpolation is performed between adjacent layers to make the corrosion area parameter change continuously in the thickness direction, thereby constructing the three-dimensional morphology of the defect; based on the interpolated continuous function, combined with the wall thickness h corresponding to each layer... k By performing integration, the volume V of the adjacent interlayer region can be reconstructed. k Complete the linear interpolation and volume V between all adjacent layers. k By summing the results, we obtain the three-dimensional morphology and total volume V of the entire corrosion defect. The formula for calculating the defect volume is: , where Δh k This represents the thickness increment corresponding to the k-th layer; The corrosion rate calculation unit is used to perform linear fitting on all wall thickness maps based on historical time series wall thickness data to calculate the corrosion rate. The remaining service calculation unit is used to calculate the remaining service based on the current minimum wall thickness, the minimum allowable wall thickness, and the corrosion rate. The graded early warning unit is used to calculate a comprehensive risk score based on a weighted comprehensive assessment of the thinning rate, corrosion allowance, and corrosion rate, and to classify alarm levels according to the score thresholds; the thinning rate is calculated based on the original wall thickness and the current minimum wall thickness, and the corrosion allowance is calculated based on the current minimum wall thickness and the allowable minimum wall thickness.

6. The online monitoring system for corrosion defects of metal pipes and equipment based on pulsed eddy currents according to claim 1, characterized in that, The system includes pre-set exclusive monitoring calibration curves for different materials such as carbon steel, stainless steel, and alloy steel.

7. The online monitoring system for corrosion defects of metal pipes and equipment based on pulsed eddy currents according to claim 1, characterized in that, The remote data transmission unit includes, but is not limited to, a Bluetooth module, a 4G module, a 5G module, an IoT communication module, and a local storage module. The local storage module is used to temporarily store the collected data when the network is interrupted and to automatically resume transmission after the network is restored.

8. The online monitoring system for corrosion defects of metal pipes and equipment based on pulsed eddy currents according to claim 1, characterized in that, The corrosion defects include, but are not limited to, wall thinning defects, body crack defects, and weld crack defects in metal pipes and equipment.

9. A method for online monitoring of corrosion defects in metal pipes and equipment based on pulsed eddy currents, applied to the online monitoring system for corrosion defects in metal pipes and equipment based on pulsed eddy currents as described in any one of claims 1 to 8, characterized in that, Includes the following steps: Step S1: System Deployment and Parameter Setting Adjust the motor running track to a suitable monitoring mode according to the geometry of the equipment to be tested and the monitoring range requirements, establish a communication connection between the pulse eddy current online monitor and the pulse eddy current surface detection sensor unit, and preset the acquisition time parameters, acquisition cycle parameters, step parameters and signal transmission and reception parameters through the pulse eddy current online monitor; Step S2: Automatic wake-up and time-sharing data acquisition For real-time monitoring, the system automatically wakes up at a preset time. The pulse eddy current online monitoring instrument controls a high-precision stepper motor to drive a multi-sensor synchronous motion module to perform a "walk-stop-collect-walk" cyclic working mode along the motor's running track. ① The multi-sensor synchronous motion module stops after moving a preset number of steps; ② Control the pulse excitation probe to emit a suitable pulse eddy current signal, and control the signal receiving probe to collect the secondary magnetic field attenuation signal; ③ After associating the collected data at the current location with the location information and collection time, temporarily store it; ④ Repeat the above steps until all preset points have been collected; For non-real-time monitoring, personnel use a handheld pulsed eddy current online monitoring instrument to move to the location of each pipe fitting equipped with a pulsed eddy current surface detection sensor unit. After connecting the monitoring instrument to the sensor unit, the equipment automatically drives the sensor unit to perform a "walk-stop-collect-walk" cycle according to preset step parameters and time-sharing working mode, sequentially completing the data collection of one or more pipe fittings. After completing the data collection of the current pipe fitting, the connection is disconnected and the equipment moves to the next pipe fitting to be tested to continue the operation. Step S3: Data Encapsulation and Remote Transmission The device automatically encrypts and encapsulates the collected data and device status information, and sends it to the data server for storage via a remote data transmission unit. Step S4: Data Processing and Multidimensional Analysis The collected data is processed and analyzed, including: ①Assess the usability of the collected data and remove invalid data with abnormal voltage attenuation characteristics; ② Extract characteristic values ​​reflecting changes in wall thickness from the valid data; ③ Perform multi-layer penetration imaging processing based on feature values ​​to generate corrosion distribution images at different depths, and then overlay them to form a three-dimensional corrosion distribution map; ④ Perform differential analysis between the corrosion image at the current moment and the baseline image at a historical moment. Using the initial acquired data as a benchmark, construct a difference matrix to generate a multi-layer penetration imaging map of corrosion change trend. ⑤ The corrosion rate is calculated by performing linear fitting based on historical time series wall thickness data; ⑥ Calculate the remaining service life of the equipment based on the current minimum wall thickness, the minimum allowable wall thickness, and the corrosion rate; ⑦ Based on a weighted comprehensive assessment of thinning rate, corrosion allowance and corrosion rate, calculate the comprehensive risk score and classify the warning level according to the score threshold; Step S5: Visualizing the Results The multi-layer penetration imaging map, corrosion change trend map, three-dimensional corrosion distribution map, corrosion rate, remaining life and early warning level information generated in step S4 are visualized through the system display module, and maintenance measures suggestions are automatically given based on historical monitoring results.

Images