Nondestructive rapid detection method for cathode protection effect of PCCP (prestressed concrete cylinder pipe) of HM water supply project

Abstract

The invention discloses a nondestructive rapid detection method for the cathode protection effect of an HM water supply project PCCP pipeline, and belongs to the technical field of buried metal pipeline corrosion protection and nondestructive detection. The method solves the problem that the measurement result is seriously distorted due to the high resistance characteristic of the PCCP concrete protection layer in the traditional potential measurement method. According to the method, a distributed optical fiber current sensing unit is adopted to continuously measure and protect current density distribution in a non-contact mode in the axial direction of a pipeline, and an actually measured current density sequence serves as a dynamic boundary condition to be input into a pre-constructed PCCP multi-layer structure current potential field coupling inversion model; and solving the coupling physical field equation set to directly invert and calculate a real polarization potential distribution map of the surface of the steel cylinder under the concrete layer, and comparing a preset protection potential standard to realize whole-line protection effect evaluation and risk section accurate positioning. According to the method, the full-line cathode protection effectiveness diagnosis and risk assessment can be quickly carried out on the PCCP of the in-service water supply project without cutting off water and excavating.

Description

Technical Field

[0001] This invention relates to the field of corrosion protection and non-destructive testing technology for buried metal pipelines. More specifically, this invention relates to a rapid non-destructive testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects. Background Technology

[0002] In existing technologies, the detection of cathodic protection effectiveness of PCCP pipelines in HM water supply projects mainly faces the following problems: First, widely used pipe-to-soil potential measurement methods, such as the closed-circuit potential method or the potential gradient method, are significantly affected by the high resistivity of the concrete protective layer of the PCCP pipeline, resulting in severe IR drop errors. This leads to the measured surface potential failing to accurately reflect the polarization potential of the internal steel cylinder, causing misjudgments of the protection status. The fundamental reason is that the steel cylinder is encased in poorly conductive concrete, making it difficult for traditional methods to penetrate this medium layer for direct measurement. The difficulty in solving this problem lies in how to obtain direct or indirect physical quantities that characterize the true electrochemical state of the steel cylinder without interrupting water supply or excavation. Second, to improve the accuracy of the assessment, some methods attempt to establish numerical models for inversion calculations. However, key parameters required for the model, such as soil resistivity and concrete conductivity, exhibit spatiotemporal variability in the engineering field. Using fixed or design parameters will lead to model distortion, making the inversion results unreliable. The difficulty lies in how to economically and conveniently obtain and update these dynamic parameters in complex non-laboratory environments to maintain the long-term accuracy of the model. Furthermore, when underprotected areas are detected, selecting the optimal repair or intervention measures becomes a challenge. Different measures, such as increasing anodes or adjusting output, have complex and global effects on the overall protection potential distribution. Traditional empirical decision-making methods cannot quantitatively assess the specific effects of each option in advance, let alone predict the potential negative interference to other sections of the pipeline, such as causing overprotection or underprotection at distant points. This makes the selection of repair options trial-and-error, potentially leading to wasted costs or new risks. In addition, structural damage to PCCP pipelines, especially prestressed wire breakage, is closely related to cathodic protection failure. However, existing nondestructive testing techniques, such as transient electromagnetic methods, mainly focus on identifying the broken wire itself, while conventional cathodic protection testing only assesses the potential. The two technologies are independent of each other, making it impossible to clearly answer the key engineering question in the inspection report: "Has the discovered broken wire caused cathodic protection failure in this area?" The difficulty in effectively correlating and analyzing the two lies in the lack of means to simultaneously acquire and couple structural and electrochemical state data under the same spatiotemporal reference. Finally, from an engineering implementation perspective, comprehensive inspection of long-distance pipelines requires high efficiency and minimal disruption to normal operation. Traditional high-precision detection methods often require densely deployed sensors or point-by-point measurements, which is time-consuming and contradicts the requirements of continuous operation in water supply projects. How to significantly shorten on-site operation time and reduce the occupation of the area above the pipeline while ensuring detection accuracy is a common challenge in practical engineering. Summary of the Invention

[0003] One object of the present invention is to solve at least the above-mentioned problems and to provide at least the advantages that will be described later.

[0004] Another objective of this invention is to provide a non-destructive and rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects. This method aims to solve the technical problem that traditional pipe-to-ground potential measurement methods suffer from severe distortion of measurement results due to the high resistivity of the concrete protective layer of PCCP pipelines, failing to accurately reflect the cathodic protection status of the internal steel cylinder. This method is used for rapid, full-line cathodic protection effectiveness diagnosis and risk assessment of in-service water supply projects' PCCP pipelines without interrupting water supply or excavation, providing direct evidence for precise pipeline maintenance and safety management.

[0005] To achieve these objectives and other advantages according to the present invention, a non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects is provided, comprising the following steps: S1. Distributed Axial Protection Current Density Measurement: Distributed fiber optic current sensing units are axially deployed along the outer wall surface or under the soil layer of the PCCP pipeline to be tested. Based on the Faraday magneto-optical effect, the distributed fiber optic current sensing units perform non-contact, uninterrupted continuous sensing of the axial circumferential magnetic field generated by the cathodic protection current of the steel cylinder inside the PCCP pipeline, thereby acquiring and recording the time-related axial protection current density data sequence continuously distributed along the pipeline axis in real time. S2. Construct a PCCP pipeline current-potential field coupled inversion model: Based on the actual structural parameters, material property parameters and environmental parameters of the PCCP pipeline, construct a multi-layer physical field finite element model including the steel cylinder, prestressed steel wire, concrete protective layer, anti-corrosion layer and soil medium; use the axial protection current density data sequence measured in step S1 as the dynamic boundary condition input of the current-potential field coupled inversion model. S3. Distributed inversion calculation of the true polarization potential of the steel cylinder: Run the current-potential field coupled inversion model, solve the coupled equations of the current field and potential field under dynamic boundary conditions, calculate the true polarization potential inside the concrete protective layer and on the outer surface of the steel cylinder in real time, and generate a true polarization potential spectrum continuously distributed along the pipeline axis. S4. Comprehensive evaluation and location of cathodic protection effect: The true polarization potential spectrum obtained by inversion in step S3 is compared point by point with the preset standard range of effective cathodic protection potential; the pipeline sections whose potential values ​​fall outside the standard range are identified and located, which are determined to be cathodic protection failure or underprotection risk sections, and a comprehensive evaluation report including the geographical location and risk level of the risk section is output.

[0006] Preferably, between steps S2 and S3, an online model calibration step is also included, specifically: At least two potential verification points with known spatial locations are set up along the PCCP pipeline to be tested. At each potential verification point, the actual polarization potential of the outer surface of the steel cylinder at that point is directly measured by a pre-embedded or minimally invasively installed embedded reference electrode. The actual polarization potential of each potential verification point is compared one by one with the calculated potential value of the corresponding position obtained by inversion from the current current density data based on the initial model parameters. Based on the deviation generated by the comparison, the dielectric electrical parameters of the region adjacent to the potential verification point in the current-potential field coupled inversion model are dynamically corrected by the optimization algorithm until the error between the calculated potential value and the measured true polarization potential is less than the set threshold, thereby obtaining a calibrated high-precision coupled inversion model. In step S3, the calibrated high-precision coupled inversion model is used to perform distributed inversion calculation of the true polarization potential of the steel cylinder.

[0007] Preferably, after step S4, the method further includes the step of: S5. Digital twin model construction: Based on the current-potential field coupling inversion model constructed in step S2 and optimized by the online model calibration step between steps S2 and S3, and combined with the current state data collected in step S1 and the evaluation report output in step S4, a benchmark digital twin model consistent with the current actual pipeline protection state is established. S6. Virtual Intervention and Effect Simulation: In the operation interface of the benchmark digital twin model, at least one proposed cathodic protection system intervention measure is defined parametrically. The intervention measures include adding auxiliary anodes, adjusting the output parameters of the potentiostat, and locally repairing the anti-corrosion layer. Based on the definition of the intervention measures, the digital twin model updates the model parameters and boundary conditions in real time, and calculates and simulates the predicted true polarization potential distribution along the entire steel cylinder of the pipeline after the intervention measures are applied. S7. Scheme Comparison and Optimization: Compare and analyze the predicted real polarization potential distribution maps of each intervention scheme obtained from simulation calculations, evaluate the effect of each scheme on eliminating the original risk section, and at the same time evaluate whether it will cause new underprotected or overprotected risk areas throughout the entire line; based on the predetermined technical and economic optimization objectives, select the recommended intervention scheme from the simulated schemes, and output the simulation effect report and key construction parameters of the scheme.

[0008] Preferably, step S1 further includes: while performing distributed axial protection current density measurement, simultaneously performing non-contact multi-frequency impedance spectrum measurement along the pipeline axis to obtain complex impedance spectrum data sequences at each measurement point on the outer wall surface of the pipeline. In step S4, the comprehensive evaluation and localization also includes: inputting the complex impedance spectrum data sequence into the pre-trained wire breakage identification model to obtain the probability and density distribution map of prestressed steel wire breakage along the pipe axis; Spatial overlay and correlation analysis were performed on the probability and density distribution of broken wires and the actual polarization potential spectrum. Based on the correlation analysis results, the following failure mode types were identified and distinguished: the underprotected risk section caused by the current shielding effect due to the breakage of the prestressed steel wire, and the underprotected risk section caused by insufficient output of the cathodic protection system.

[0009] Preferably, in step S4, after obtaining the broken wire probability and density distribution map and before performing spatial overlay and correlation analysis, a model on-site verification and adaptive optimization step is also included, specifically: Based on the probability and density distribution of broken wires and the actual polarization potential spectrum, at least two representative field verification points were selected. For each on-site verification point, the actual physical state verification results of the prestressed steel wire at that point are obtained through minimally invasive testing or internal video inspection. The actual physical state verification results are compared with the initial diagnostic results output by the broken wire identification model for that point to generate a model accuracy evaluation report. If the comparison error exceeds the preset threshold, the verification results of the on-site verification points and their corresponding real physical states will be combined with the original complex impedance spectrum data to form an incremental training sample set, and the pre-trained broken wire identification model will be subjected to on-site incremental learning and parameter fine-tuning. Using the finely tuned and optimized broken wire identification model, the complex impedance spectrum data sequence is reprocessed to generate an updated broken wire probability and density distribution map that is more suitable for the specific conditions of the current pipeline under test, which is then used for subsequent spatial overlay and correlation analysis.

[0010] Preferably, step S1 specifically includes the following sub-steps: S1a. Rapid vehicle-mounted magnetic anomaly survey: On the ground directly above the PCCP pipeline to be tested, drive a testing vehicle equipped with a multi-channel high-precision fluxgate sensor array, and drive at a constant speed of no less than 20 kilometers per hour along the pipeline route; the sensor array measures the vertical and horizontal components of the surface magnetic field generated by the cathodic protection current of the pipeline in real time and synchronously, forming a surface magnetic field intensity spectrum continuously distributed along the pipeline axis. S1b, Intelligent location of abnormal protection current density sections: Based on the surface magnetic field strength map, the distribution trend of protection current density along the pipeline axis is quickly calculated and identified through the magnetic field-current inversion algorithm, and suspected abnormal sections with significant abrupt changes, attenuation or distortion of current density are located. The length range of suspected abnormal sections is 50-200 meters. S1c, Optimized deployment of distributed optical fiber sensing units: Distributed optical fiber current sensing units are precisely deployed only within each suspected abnormal section located in step S1b and within a 50-meter extension range upstream and downstream; For the parts of the pipeline that are not identified as suspected abnormal sections, distributed optical fiber current sensing units are not deployed, and only the surface magnetic field survey data obtained in step S1a is retained as background reference.

[0011] Preferably, after completing the comprehensive evaluation and positioning in step S4, the process also includes a data fusion and unified spatiotemporal archive construction step, specifically: Establish a unique digital pipeline archive associated with the PCCP pipeline under test: All source and derived data generated during this inspection, including the spatial geographic coordinates of each measurement point, acquisition timestamp, original value of axial protection current density, original data of multi-frequency impedance spectrum, surface magnetic field data, true polarization potential obtained from inversion calculation, wire breakage diagnosis results, model calibration parameters, simulation prediction data, and the aforementioned comprehensive evaluation report, are synchronously stored in the digital pipeline archive with a unified data structure; the digital pipeline archive supports multi-dimensional data association queries, comparative analysis, and visualization based on spatial location and time.

[0012] Preferably, after the digital pipeline archive is constructed, it also includes a long-term trend analysis and early warning step, specifically: In the digital pipeline archive, the true polarization potential and broken wire density data obtained from each detection at the same spatial location point are extracted in time sequence. Based on time series data, a potential decay trend prediction model and a wire breakage development rate prediction model were constructed respectively. When the predictive model indicates that the potential value of a specific section will drop below the protection standard threshold within a predetermined period, or that the broken wire density will exceed the safety threshold, the system automatically generates a warning alert and identifies the key time points in the risk evolution.

[0013] Preferably, for the underprotected risk sections identified in step S4, a failure root cause contribution quantification analysis step is also included, specifically: For the underprotected risk section, a multi-scenario simulation comparison model is constructed based on the current-potential field coupling inversion model; The multi-scenario simulation comparison model simulates and calculates the protection potential distribution under the following conditions: only the current diagnostic wire breakage state exists; only the measured environmental medium parameters exist; and other single assumed failure factors exist. By comparing the differences in the degree of agreement between each simulation result and the measured potential distribution obtained from the inversion in step S3, the independent contribution and coupling contribution of wire breakage factors, environmental medium factors and other factors to the current potential decay result are quantitatively evaluated, and a root cause quantitative analysis report is generated.

[0014] Preferably, in step S1b, the magnetic field-current inversion algorithm also integrates known geological survey data and historical excavation records along the pipeline route; The intelligent positioning process uses the fused data to filter out environmental interference and correct geological stratification in the magnetic field intensity map, and weights the credibility of suspected abnormal sections located by the algorithm based on the spatial location of known defects or interference points in historical records. Output a list of suspected abnormal sections with different confidence levels and corresponding decision recommendations. The decision recommendations include conducting a detailed fiber optic survey immediately for high-confidence sections, suggesting supplementing medium-confidence sections with close-range magnetic surveys or ground-penetrating radar scans, and marking low-confidence sections as long-term monitoring points of concern.

[0015] The present invention has at least the following beneficial effects: 1. The non-destructive rapid detection method for the cathodic protection effect of PCCP pipelines in HM water supply projects of the present invention effectively overcomes the IR drop error in the traditional pipe-to-ground potential method by measuring current density and combining it with physical field coupling inversion. For the first time, it obtains the true polarization potential distribution on the surface of the steel cylinder under the concrete layer without contacting the steel cylinder, fundamentally solving the core problem of measurement distortion.

[0016] 2. The non-destructive rapid detection method for the cathodic protection effect of PCCP pipelines in HM water supply projects of the present invention performs online calibration of the inversion model through field verification points and incremental learning of the intelligent diagnostic model by using a small amount of excavation verification, so that the entire detection system can adapt to the dynamic changes of specific engineering environment and ensure the reliability of the results of the method in different projects and long-term use.

[0017] 3. The non-destructive rapid detection method for the cathodic protection effect of PCCP pipelines in HM water supply projects of the present invention is based on a precise digital twin model to simulate and rehearse intervention measures, and uses historical archive data for trend prediction and early warning. This transforms the operation and maintenance mode from reactive post-maintenance to proactive pre-maintenance optimization, significantly improving the scientificity, economy and safety of maintenance decisions.

[0018] 4. The non-destructive rapid detection method for the cathodic protection effect of PCCP pipelines in HM water supply projects of the present invention, through synchronous measurement and data correlation analysis, clearly distinguishes the failure modes caused by broken wire shielding, insufficient system output or a combination of both, and upgrades the isolated defect list into a mechanistic diagnostic report containing causal relationships, guiding maintenance resources to be accurately directed to the root cause.

[0019] 5. The non-destructive rapid detection method for the cathodic protection effect of PCCP pipelines in HM water supply projects of the present invention uses vehicle-mounted magnetic measurement to quickly locate suspected abnormal sections, thereby concentrating high-cost and high-precision detection resources on key areas. While ensuring diagnostic depth, it significantly improves the operational efficiency of long-distance pipeline full-line inspection and reduces overall costs.

[0020] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Detailed Implementation

[0021] The present invention will now be described in further detail so that those skilled in the art can implement it based on the description.

[0022] This invention provides a non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects, which includes the following steps: S1. Distributed Axial Protection Current Density Measurement: Distributed fiber optic current sensing units are axially deployed along the outer wall surface or under the soil layer of the PCCP pipeline to be tested. Based on the Faraday magneto-optical effect, the distributed fiber optic current sensing units perform non-contact, uninterrupted continuous sensing of the axial circumferential magnetic field generated by the cathodic protection current of the steel cylinder inside the PCCP pipeline, thereby acquiring and recording the time-related axial protection current density data sequence continuously distributed along the pipeline axis in real time. S2. Construct a PCCP pipeline current-potential field coupled inversion model: Based on the actual structural parameters, material property parameters and environmental parameters of the PCCP pipeline, construct a multi-layer physical field finite element model including the steel cylinder, prestressed steel wire, concrete protective layer, anti-corrosion layer and soil medium; use the axial protection current density data sequence measured in step S1 as the dynamic boundary condition input of the current-potential field coupled inversion model. S3. Distributed inversion calculation of the true polarization potential of the steel cylinder: Run the current-potential field coupled inversion model, solve the coupled equations of the current field and potential field under dynamic boundary conditions, calculate the true polarization potential inside the concrete protective layer and on the outer surface of the steel cylinder in real time, and generate a true polarization potential spectrum continuously distributed along the pipeline axis. S4. Comprehensive evaluation and location of cathodic protection effectiveness: The true polarization potential spectrum obtained from step S3 is compared point by point with the preset standard range of effective cathodic protection potential (according to GB / T 21448-2017 standard, the effective range is -0.85V to -1.2V); the pipeline sections whose potential values ​​fall outside the standard range are identified and located, which are determined to be cathodic protection failure or underprotection risk sections, and a comprehensive evaluation report including the geographical location and risk level of the risk section is output.

[0023] In the above technical solution, the preliminary preparation and equipment deployment are as follows: First, a site survey is conducted to obtain the precise route, burial depth, soil resistivity, and specific structural parameters of the PCCP pipeline, including the steel cylinder diameter, concrete protective layer thickness, and anti-corrosion layer type. For example, in this instance, the pipeline burial depth is two meters, and the standard thickness of the concrete protective layer is 50 millimeters. Subsequently, distributed fiber optic current sensing units are axially deployed along the ground surface directly above the pipeline or by slightly excavating the overburden layer to the surface of the pipeline's outer anti-corrosion layer. The core of this distributed fiber optic current sensing unit is a single-mode communication optical fiber (core diameter 9 μm, cladding diameter 125 μm), whose surface is coated with a special magnetic material coating (thickness 5~10 μm) to enhance the sensitivity of the Faraday magneto-optical effect. The fiber spatial resolution is ≥1m, and the current density measurement range is 0~100mA / m. 2 The measurement accuracy is ±1% FS. The optical fiber is laid close to the outer wall of the pipe, with a fixed point every 100 meters to ensure it is parallel to the pipe axis and forms a continuous sensing path. The starting and ending ends of the optical fiber are connected to a high-precision distributed optical fiber sensor demodulator. This instrument emits probe light and analyzes the polarization state changes of backscattered light caused by the pipe's circumferential magnetic field. The final output data is the axial protection current density value in milliamperes per square meter, with a spatial resolution of 1 meter along the optical fiber path.

[0024] Data Acquisition and Model Building: Under normal power-on operation of the cathodic protection system, the sensor demodulator is activated for continuous measurement. The measurement duration is recommended to cover at least one complete power supply cycle, such as 24 hours. The instrument will collect and record dynamic current density data distributed along the entire pipeline in real time. This data directly reflects the distribution intensity of the protective current output from the cathodic protection station and flowing along the steel cylinder surface along the pipeline axis. Simultaneously, a current-potential field coupled inversion model of this pipeline section is constructed on a computer. The model is based on the finite element method, with its geometry consistent with the actual pipeline. Material properties are set according to previously prepared data: for example, the steel cylinder conductivity is set to 5 million Siemens per meter, the concrete resistivity to 100 ohm-meters, and the soil resistivity to 50 ohm-meters. The core physical fields of the model are the DC current field and the electrostatic field, coupled through the constitutive relation of dielectric conductivity. The boundary conditions of the model are crucial. The current injection boundary along the pipeline axis is no longer the ideal value assumed by traditional methods, but rather the dynamic current density data sequence obtained in the previous step is input as a time-varying non-uniform boundary condition to the corresponding pipeline outer wall location in the model. The soil surface is set as the zero potential point.

[0025] True Potential Inversion Calculation and Effect Evaluation: The pre-constructed coupled inversion model is run for solution. The model calculation software calculates the distribution of current lines and equipotential lines throughout the entire field under given measured current density boundary conditions. Through calculation, the potential value on the outer surface of the steel cylinder encased in concrete can be directly output. This potential value is the true polarization potential of the steel cylinder after removing the effects of IR drop caused by concrete pressure drop, soil pressure drop, etc. The calculation result is a true potential spectrum corresponding one-to-one with the measured current density points along the entire pipeline, with values ​​typically ranging from -0.8V to -1.2V. Finally, a comprehensive evaluation of the protection effect is performed. The true polarization potential data obtained from the inversion is compared with the cathodic protection effectiveness criteria. For example, according to relevant standards, if the true polarization potential of the steel cylinder under energized conditions remains in a range more negative than -0.85V for a long period, the protection is considered effective. Through automatic scanning by the software, all pipeline locations with potential values ​​positive than -0.85V, such as those as low as -0.75V, can be quickly identified. The continuous pipe sections formed by these points are identified as underprotected risk sections. The system ultimately generates a graphic report, highlighting these risk sections on a geographic information system map and marking their start and end station numbers, length, and degree of potential deviation, providing direct basis for maintenance decisions.

[0026] Traditional methods use a copper sulfate reference electrode to measure the "pipe-to-ground potential" at the ground surface. This potential is the total pressure drop generated by the protective current flowing through the soil and concrete layer, which includes a significant IR drop error. In contrast, this invention directly measures the protective current density distributed along the pipeline axis, which is a field source quantity independent of the resistivity of the medium, thus avoiding the interference of IR drop at its source.

[0027] Traditional methods can only obtain a vague and distorted potential value at the ground surface, failing to reveal the true electrochemical state of the steel cylinder beneath the concrete layer. This invention, through physical model inversion, achieves for the first time a precise mapping from the externally measurable state to the internal true state, directly outputting the true polarization potential of the steel cylinder surface. This is the gold standard for evaluating the effectiveness of cathodic protection.

[0028] Traditional methods rely on measurements from a finite number of discrete points for coarse judgments, easily missing small underprotected areas. This invention provides a continuous, full-line potential spectrum, enabling meter-level accuracy in locating risk zones, providing a comprehensive and precise assessment.

[0029] This technical solution creatively introduces a new technical approach that combines "distributed current density sensing" with "physical field coupling inversion". For the first time, it achieves non-destructive, rapid, quantitative, and direct diagnosis of the true effect of cathodic protection inside the complex multi-layered structure of PCCP pipelines without interrupting operation or excavation. It fundamentally solves the technical problems of measurement distortion, inaccurate evaluation, and low efficiency of traditional methods.

[0030] In another technical solution, an online model calibration step is also included between steps S2 and S3, specifically: At least two potential verification points with known spatial locations are set up along the PCCP pipeline to be tested. At each potential verification point, the actual polarization potential of the outer surface of the steel cylinder at that point is directly measured by a pre-embedded or minimally invasively installed embedded reference electrode. The actual polarization potential of each potential verification point is compared one by one with the calculated potential value of the corresponding position obtained by inversion from the current current density data based on the initial model parameters. Based on the deviation generated by the comparison, the dielectric electrical parameters within a 50-meter range adjacent to the potential verification point in the current-potential field coupled inversion model are dynamically corrected by the least squares optimization algorithm until the error between the calculated potential value and the measured true polarization potential is less than the set threshold, thereby obtaining a calibrated high-precision coupled inversion model. In step S3, the calibrated high-precision coupled inversion model is used to perform distributed inversion calculation of the true polarization potential of the steel cylinder.

[0031] Preferably, the deviation generated by the comparison is dynamically corrected for the dielectric electrical parameters using a least squares optimization algorithm. The specific process is as follows: Let the model parameter vector to be corrected be... i (For example, the concrete resistivity covering the area adjacent to each verification point). Define the objective function F( i ) = Σ j [φ cal ( i , x j ) - φ meas (x j )] 2 , where φ cal For the model in the parameter vector i Below, at verification point x j The calculated potential value at φ meas These are the measured potential values ​​at the corresponding verification points. The online calibration process of the model is essentially a problem of minimizing the objective function described above, i.e. .

[0032] This invention employs the Levenberg-Marquardt nonlinear least squares optimizer to solve the problem. This algorithm can efficiently handle such inversion problems; its core logic involves iteratively updating the parameter vector. i until the objective function value F( i The algorithm converges to a preset threshold (e.g., the corresponding potential error is less than 10 mV). In this algorithm, the initial value of the damping factor is set to 0.01, and the maximum number of iterations is set to 100.

[0033] In the above technical solution, an online model calibration step is added to ensure the long-term reliability of the inversion results.

[0034] Establishment and Direct Measurement of Potential Verification Points: While implementing distributed fiber optic deployment, potential verification points need to be established at key locations along the pipeline. These points should be representative, for example, one every 1 kilometer, with additional points added near abrupt changes in geological conditions and suspected defects. The specific implementation for each verification point is as follows: A minimal excavation is carried out directly above the selected pipeline until the concrete protective layer is exposed. Subsequently, using minimally invasive drilling technology with an 8mm diameter diamond drill bit, a 50mm deep guide hole is drilled into the concrete protective layer. This depth ensures that the internal prestressed steel wires are not damaged. Then, a specially designed long-lasting embedded silver chloride reference electrode is implanted. The electrode's tip is a porous ceramic plug, which, through the injection of a special conductive gel, forms a stable electrochemical contact with the steel cylinder surface. The electrode lead is led out to a waterproof junction box on the ground. Special low-resistivity filler is used during backfilling to ensure stable contact between the electrode and the surrounding medium. The entire installation process has a negligible impact on the pipeline structure. During each full-line current density measurement, a high-impedance voltmeter is used simultaneously to measure the potential of the embedded electrode relative to the remote copper-copper sulfate reference electrode at each verification point. This reading is the actual polarization potential of the steel cylinder surface at that point, with an accuracy of ±5 millivolts.

[0035] The adaptive calibration process of the model: After completing the current density data acquisition and verification point potential measurement, the model calibration phase begins. First, the initial current-potential field coupled inversion model (established based on the average parameters of the design drawings and survey) is used for the first inversion calculation based on the current current density data to obtain the calculated potential values ​​at each point along the entire line, and the calculated potential values ​​corresponding to each potential verification point are extracted. Next, the calibration algorithm is activated. For example, the measured true polarization potential of the first verification point (located at chainage K1+000) of -0.910 V is compared with the model-calculated potential of -0.880 V at the same point, revealing a positive deviation of 30 mV. The calibration algorithm then starts, aiming to minimize the potential deviation at this point, and automatically adjusts the key parameters in the model of the approximately 50-meter section of pipe near the verification point. For example, the effective resistivity of the concrete in this area is gradually reduced from the initial setting of 100 ohm-meters to approximately 85 ohm-meters. After completing the local parameter correction for the first point, the algorithm processes the second and third verification points in sequence. The calibration is complete when the absolute error between the calculated and measured potential values ​​at all verification points is less than the preset threshold of 10 millivolts, generating a calibrated high-precision coupled inversion model. This model not only integrates the current density distribution information of the entire pipeline but also embeds the actual potential constraints at multiple locations, and its physical field characterization is closer to the current actual state of the pipeline.

[0036] High-precision inversion and evaluation based on the calibration model: Finally, using this calibrated model, the distributed inversion calculation of the true polarization potential along the entire line is re-executed. Since the model parameters have been corrected using real field data, the confidence level of the full-line potential spectrum calculated in this inversion is significantly improved. Based on this high-confidence spectrum, the system then performs the final evaluation of the cathodic protection effect and locates risk sections. The output evaluation report will include a description of the model calibration accuracy, such as "This evaluation has calibrated the model based on measured data from eight verification points along the line, with an average calibration error of 7 millivolts at each point," thus greatly enhancing the authority of the evaluation conclusions and their engineering guidance value.

[0037] Compared to existing traditional or model-based methods that lack a calibration mechanism, this implementation completely breaks this assumption by introducing field potential verification points and a dynamic calibration cycle. The calibration process enables the model to "sense" environmental changes (such as soil drying or concrete wetting) and "self-update," a capability completely absent in traditional static models. Traditional methods improve accuracy by increasing the precision of individual measuring instruments or increasing the number of measurement points, but cannot address systemic model errors. This implementation creates a closed-loop feedback accuracy control mechanism. It uses a small number of absolute truth measurements as "anchor points" to guide and correct the entire computational model, thereby improving the accuracy of tens of thousands of calculation points.

[0038] This implementation presents an embedded, automated, and low-cost self-maintaining system for model accuracy. It can stably serve engineering practice over a long period, providing reliable decision-making data, thus solving the reliability challenge of advanced algorithms when transitioning from the laboratory to complex real-world engineering environments.

[0039] In another technical solution, after step S4, the following step is also included: S5. Digital twin model construction: Based on the current-potential field coupling inversion model constructed in step S2 and optimized by the online model calibration step between steps S2 and S3, and combined with the current state data collected in step S1 and the evaluation report output in step S4, a benchmark digital twin model consistent with the current actual pipeline protection state is established. S6. Virtual Intervention and Effect Simulation: In the operation interface of the benchmark digital twin model, at least one proposed cathodic protection system intervention measure is defined parametrically. The intervention measures include adding auxiliary anodes, adjusting the output parameters of the potentiostat, and locally repairing the anti-corrosion layer. Based on the definition of the intervention measures, the digital twin model updates the model parameters and boundary conditions in real time, and calculates and simulates the predicted true polarization potential distribution along the entire steel cylinder of the pipeline after the intervention measures are applied. S7. Scheme Comparison and Optimization: Compare and analyze the predicted real polarization potential distribution maps of each intervention scheme obtained from simulation calculations, evaluate the effect of each scheme on eliminating the original risk section, and at the same time evaluate whether it will cause new underprotected or overprotected risk areas throughout the entire line; based on the predetermined technical and economic optimization objectives, select the recommended intervention scheme from the simulated schemes, and output the simulation effect report and key construction parameters of the scheme.

[0040] In the above technical solution, the construction of the benchmark digital twin model is as follows: Simultaneously with generating the cathodic protection effectiveness assessment report, the detection system software automatically launches a twin model construction module. This module uses a high-precision current-potential field coupling inversion model calibrated at eight field potential verification points as its core digital kernel. Simultaneously, the module binds and associates all input and output data of this detection with this kernel. This data includes: the precise three-dimensional spatial coordinates of the pipeline, the measured current density value per meter along the pipeline, the calibrated full-line true polarization potential spectrum, and the identified risk section locations (e.g., the potential of the section from station K3+200 to K3+500 is -0.78V). All this data together constitutes a benchmark digital twin model reflecting the current instantaneous true state of the pipeline. This model is visualized as a three-dimensional pipeline in the software interface, with its color changing gradient according to the potential value. Risk sections are highlighted in flashing red, and interactive operations such as scaling and rotation are supported.

[0041] Virtual Intervention and Dynamic Effect Simulation: When engineers need to develop maintenance plans for the aforementioned risk areas, they can operate on this twin model. The software provides a "Virtual Intervention Toolbox." For example, an engineer's first plan might be to "add a temporary magnesium alloy sacrificial anode at point K3+350 in the midpoint of the risk area." The engineer selects "Add Anode" in the toolbox and clicks to place a virtual anode icon at position K3+350 in the model's 3D view. Then, in the pop-up parameter panel, the anode's properties are set: output current is set to 5 amps, grounding resistance to 1 ohm, and expected service life to 2 years. After setting, click "Simulate." The software's core solver will immediately start from the current baseline state, using "adding a point source with an output current of 5 amps at the specified location" as the new boundary condition, and resolve the entire coupled physical field. Approximately 10 seconds later, the simulation calculation is complete. The model view is globally updated; the original red risk area color begins to change to a dark green representing good protection, but it may not be completely eliminated. Meanwhile, the software automatically generated a post-simulation predicted potential distribution map and provided key data: the average potential in the original risk section increased from -0.78V to -0.83V, but approximately 50 meters of pipe section still had a potential above -0.82V, not fully meeting the standard. Furthermore, the simulation report indicated that near the far end K8+000, the potential changed from -0.90V to -0.93V, showing a slight tendency towards overprotection.

[0042] Multi-scheme comparison and optimization decision-making: Based on the simulation results of the first scheme, engineers can quickly iterate and try other schemes. For example, the second scheme: "Install a 3-amp anode at K3+300 and K3+400 respectively." Simulation results show that this scheme can uniformly raise the potential of the entire risk section to above -0.86V, with minimal impact on other areas of the entire line and no risk of overprotection. The third scheme: "Do not add anodes, only increase the output voltage of the potentiostat at the first station by 0.1V." Simulation results show that although the potential of the risk section can be raised to -0.84V, it will cause the potential at the end of the pipeline to be too negative to -1.05V, posing a risk of coating peeling. The software provides a scheme comparison view, displaying the simulation results of the three schemes side by side, and quantifying them from three dimensions: "improvement degree of the target section," "negative effect of the global impact," and "estimated cost." Assuming that the predetermined optimization goal is "to achieve the target potential of the entire line at the lowest cost and without overprotection points," the algorithm will recommend the second scheme as the recommended intervention scheme. Finally, the system outputs a detailed "Simulation Effect Report and Construction Guidelines", which clearly states that the "dual anode local reinforcement scheme" is recommended. The precise anode installation location, required output current, expected effect and construction precautions are all clearly shown, providing the maintenance team with a precise "action map".

[0043] Compared to traditional maintenance decisions based on experience and standard manuals, this implementation method represents a significant upgrade from "qualitative experience" to "quantitative simulation." Traditional methods rely on engineer experience, referring to standard manuals to select roughly equivalent solutions (such as "adding an anode every kilometer"). This implementation method, however, uses a precise digital twin of the current pipeline for quantitative simulation. It can answer precise questions such as "at which station should this anode be installed, and what current should it output for optimal results?", elevating decision-making from a vague, experience-based level to a precise numerical optimization level. Before traditional construction, no one can definitively know whether a particular solution will trigger remote side effects. This implementation method, through global coupling simulation, reveals all potential chain reactions (such as overprotection) in advance in virtual space, enabling engineers to proactively avoid risks and achieving a shift from "risky construction" to "safe pre-construction."

[0044] Existing technologies primarily demonstrate economic value indirectly by preventing accidents. This implementation method, however, directly generates significant economic benefits. Through simulation and comparison, ineffective or harmful construction can be avoided, and the lowest-cost and most effective solution can be selected. For example, simulation may reveal that the original plan to add ten anodes along the entire line actually only requires two precisely positioned anodes to meet the standard, saving substantial material and construction costs.

[0045] In another technical solution, step S1 further includes: while performing distributed axial protection current density measurement, simultaneously performing non-contact multi-frequency impedance spectrum measurement along the pipeline axis to obtain complex impedance spectrum data sequences at each measurement point on the outer wall surface of the pipeline. In step S4, the comprehensive evaluation and localization also includes: inputting the complex impedance spectrum data sequence into a pre-trained convolutional neural network (CNN) wire breakage identification model. This model has been trained with more than 5,000 sets of 'impedance spectrum-excavation verification' data, and the wire breakage identification accuracy is ≥95%. The model obtains the probability and density distribution map of prestressed steel wire breakage along the pipeline axis. Spatial overlay and correlation analysis were performed on the probability and density distribution of broken wires and the actual polarization potential spectrum. Based on the correlation analysis results, the following failure mode types were identified and distinguished: the underprotected risk section caused by the current shielding effect due to the breakage of the prestressed steel wire, and the underprotected risk section caused by insufficient output of the cathodic protection system.

[0046] As a preferred embodiment, the pre-trained broken wire recognition model described in step S4 is constructed and trained using the following method: Model Structure: This model is a one-dimensional convolutional neural network, and its specific structure is as follows: Input layer: receiving dimension is [n] frequencies A vector of [× 2], where n frequencies The number of characteristic frequency points of the impedance spectrum (e.g., 10) is represented by 2, where 2 represents the impedance amplitude and logarithmically converted phase angle at each frequency point.

[0047] Convolution and Pooling Module: Contains three repeating units, each consisting of a one-dimensional convolutional layer (with a kernel size of 3, and the number of filters being 32, 64, and 128 respectively, using the ReLU activation function) followed by a max pooling layer (pooling size 2).

[0048] Flattening layer and fully connected layer: After flattening the output of the convolution module, it is connected to two parallel fully connected branches.

[0049] Branch 1 (Filament breakage probability): consists of a fully connected layer with 128 neurons (ReLU activation), a Dropout layer (dropout rate 0.5), and a fully connected layer with 1 neuron (Sigmoid activation), outputting a fibrillation breakage probability between 0 and 1.

[0050] Branch 2 (Filament Density): consisting of a fully connected layer with 128 neurons (ReLU activation), a Dropout layer (dropout rate 0.5), and a fully connected layer with 1 neuron (linear activation), outputting an estimated value of the fragment density (fibers / meter).

[0051] Model training: Data preparation: Collect over 5000 data pairs from historical engineering projects. Each data pair includes a complex impedance spectrum data vector and a corresponding excavation-verified label (containing a Boolean "whether the wire is broken" and a continuous "wire breakage density"). Randomly divide the dataset into training, validation, and test sets in a 7:2:1 ratio.

[0052] Preprocessing: The amplitude of the input impedance spectrum is normalized by Z-score, and the phase angle is normalized after logarithmic transformation.

[0053] Loss function and optimization: A multi-task loss function L = λ1 × L is adopted. prob + λ2× L density L prob For binary cross-entropy loss, L density The mean squared error loss is used, and λ1 and λ2 are balancing weights, with values ​​of 1.0 and 0.5 respectively. The optimizer used is Adam, with an initial learning rate of 0.001 and a batch size of 32.

[0054] Training process: Train on the training set for a maximum of 200 epochs. When the loss on the validation set no longer decreases for 10 consecutive epochs, terminate the training early and save the model parameters with the best performance on the validation set.

[0055] Performance evaluation: On the independent test set, the model achieved an accuracy of 95.2% in classifying broken wires (yes / no) and a mean absolute error of 0.35 wires / meter in estimating broken wire density.

[0056] When performing in-situ incremental learning, the parameters of all convolutional layers in the above network are frozen, and only the parameters of the last two layers of the two fully connected branches are fine-tuned. A small number of new samples consisting of in-situ validation points are used for training in 50 batches at a smaller learning rate (e.g., 0.0001) to prevent overfitting and quickly adapt to the in-situ pipeline features.

[0057] In the above technical solution, this embodiment continues the inspection task of the ten-kilometer PCCP pipe section, and assumes that the fiber optic deployment and initial setup have been completed according to the basic method.

[0058] Implementation of synchronous dual-modal data acquisition: Simultaneously with distributed fiber optic current density measurements (e.g., while the inspection vehicle is traveling along the pipeline at a speed of five kilometers per hour), multi-frequency impedance spectroscopy measurements are initiated. The inspection vehicle carries an impedance analyzer connected to a non-contact electromagnetic coupling probe. This probe is mounted at the end of a retractable robotic arm, which automatically positions the probe to a height of approximately 0.3 meters above the ground directly above the pipeline, or through a guide wheel mechanism, brings it into slight contact with the exposed corrosion-resistant surface of the pipeline.

[0059] The impedance analyzer operates in frequency sweep mode. At each measurement point (typically consistent with the spatial resolution of fiber optic measurements, set at one point per meter), a sequence of AC excitation signals containing ten characteristic frequencies (frequency range 100Hz–100kHz) is applied to the pipeline. This frequency range is selected based on the electromagnetic response characteristics of the prestressed steel wires in the PCCP pipeline, effectively distinguishing between broken wire signals and interference signals from the concrete medium. The probe receives the response signals, measures and records the impedance amplitude and phase angle at each frequency point, thereby generating the complex impedance spectrum data for that point. This process is strictly synchronized with the current density measurement of the fiber optic cable, with the same spatial positioning system recording the precise station number of each data point. For example, at station number K5+000, the system simultaneously records the real-time axial current density of 15 mA / m², as well as the corresponding complex impedance data at the ten frequencies.

[0060] Application of the Broken Wire Identification Model and Spectrum Generation: After the field data collection was completed, the indoor data processing phase began. The complex impedance spectrum data sequence of 10,000 points collected along the entire line was input into a pre-trained convolutional neural network (CNN) broken wire identification model. This model was trained on over 5,000 sets of historical 'impedance spectrum excavation verification' data, achieving a broken wire identification accuracy of ≥95%. The model input consisted of the amplitude and phase angle characteristics of the complex impedance spectrum, and the output consisted of estimated broken wire probability and density values. This model is a convolutional neural network trained using over 5,000 sets of historical PCCP pipeline 'impedance spectrum-excavation verification' data pairs, achieving a broken wire identification accuracy exceeding 95% after training. The model analyzed the characteristics of the impedance spectrum point by point, particularly the phase angle in the low-frequency band and the amplitude frequency response curve shape in the mid-frequency band, outputting three key diagnostic results: the probability of prestressed steel wire breakage at that point (expressed as a percentage, e.g., 80%), the broken wire density (expressed as the number of broken wires per meter, e.g., three per meter), and the local equivalent electrical parameters of the concrete at that point calculated based on high-frequency data. The software integrates these results to generate three curves continuously distributed along the pipeline axis, forming a spectrum of wire breakage probability and density. For example, the spectrum may show that in the section from station K6+100 to K6+300, the wire breakage probability is consistently higher than 90%, and the peak wire breakage density reaches five wires per meter.

[0061] Source data correlation analysis and precise failure mode differentiation: Finally, a core correlation analysis is performed for comprehensive evaluation and localization. The evaluation software generates two core maps: one is the true polarization potential map calculated by the inversion model, and the other is the wire breakage probability and density distribution map output by the deep learning model. The software precisely overlays these two maps based on spatial stationing and initiates the correlation analysis algorithm. The algorithm first identifies all underprotected risk sections with substandard potentials. Then, for each underprotected risk section, the algorithm examines the distribution of wire breaks in its spatial location. Specifically, it differentiates them as follows: Failure Section Dominated by Wire Breakage: In a certain underprotected risk section (e.g., K6+100 to K6+300), the potential spectrum shows a sharp drop in potential from the normal -0.9V to -0.75V. Simultaneously, the superimposed wire breakage spectrum shows a high density of broken wires in this section (e.g., an average of four broken wires per meter). Further analysis of the original current density data for this section reveals a significant dip in current density. In this case, the software determines that the failure in this section is primarily due to the physical shielding effect caused by the broken wires, severely hindering the uniform distribution of the protective current. Failure Section Due to Insufficient System Output: In another underprotected risk section (e.g., K1+000 to K1+200), the potential is substandard, but the wire breakage spectrum shows a breakage probability of less than 5% and an extremely low density. Current density data also shows an overall low level. The software determines that the failure in this section is mainly due to insufficient output from the cathodic protection station or overall weakened protection caused by a harsh soil environment. Composite failure section: In some sections, the software may identify that moderate wire breakage and moderate potential decay and current density decrease coexist, which is then marked as composite failure.

[0062] The final comprehensive assessment report not only lists the risk areas but also clearly marks the dominant failure mode type for each risk area (e.g., "fractured wire current shielding type"), along with relevant data snippets as the basis for judgment. This provides a direct and reliable decision-making basis for subsequent targeted maintenance measures (such as localized enhanced protection for the fractured wire area or adjustments to the overall output for systemic problems).

[0063] Compared to existing single transient electromagnetic method (TEM) wire breakage detection technology, this implementation method performs mandatory spatiotemporal synchronization and correlation analysis between wire breakage information and the core state parameters (true potential, current density) of the cathodic protection system. The output is a "failure mechanism diagnosis report," which not only indicates "where the problem is," but also clarifies "whether and how the problem led to the failure of the corrosion protection function," resulting in a fundamentally different diagnostic depth. The data value is multiplied: In the transient electromagnetic method, wire breakage data is isolated. In this implementation method, the engineering value of wire breakage data is greatly enhanced through correlation with electrochemical data. For example, an isolated, low-density wire breakage point, if it does not cause potential decay, can be assessed as low-risk; while a high-density wire breakage area, if it causes severe potential decay, is classified as high-risk and prioritized. This enables refined risk classification management of structural defects, which is impossible with a single detection technology.

[0064] This implementation method, through a synchronous measurement protocol and an AI-based multi-source data automatic correlation analysis algorithm, automatically completes the causal correlation diagnosis of "structural damage-electrochemical state" in a single operation at the engineering site. This is not merely a detection method, but an advanced "failure root cause analysis" system. It directly tackles the core challenge that has long plagued PCCP pipeline safety management: how to accurately determine the actual hazard of structural damage, thereby precisely directing valuable maintenance resources to the most dangerous and urgently needed pipe sections, achieving a leap from "symptom-based" to "mechanism-based" pipeline safety operation and maintenance.

[0065] In another technical solution, in step S4, after obtaining the broken wire probability and density distribution map and before performing spatial overlay and correlation analysis, a model field verification and adaptive optimization step is also included, specifically: Based on the probability and density distribution of broken wires and the actual polarization potential spectrum, at least two representative field verification points were selected. For each on-site verification point, the actual physical state verification results of the prestressed steel wire at that point are obtained through minimally invasive testing or internal video inspection. The actual physical state verification results are compared with the initial diagnostic results output by the broken wire identification model for that point to generate a model accuracy evaluation report. If the comparison error exceeds the preset threshold, the verification results of the on-site verification points and their corresponding real physical states will be combined with the original complex impedance spectrum data to form an incremental training sample set, and the pre-trained broken wire identification model will be subjected to on-site incremental learning and parameter fine-tuning. Using the finely tuned and optimized broken wire identification model, the complex impedance spectrum data sequence is reprocessed to generate an updated broken wire probability and density distribution map that is more suitable for the specific conditions of the current pipeline under test, which is then used for subsequent spatial overlay and correlation analysis.

[0066] In the above technical solution, this embodiment continues the detection of a 10-kilometer PCCP pipe section, assuming that the fiber current density measurement and multi-frequency impedance spectrum measurement have been completed simultaneously, and a preliminary fiber breakage probability distribution map and a true polarization potential map have been generated.

[0067] Intelligent Selection and Verification of Representative Field Verification Points: After completing the initial diagnosis, the system software activates the intelligent selection algorithm for verification points. This algorithm comprehensively considers the wire breakage probability, potential level, and spatial distribution, automatically recommending five field verification points according to the following principles: Highest Risk Point: Select one point with the highest wire breakage probability (e.g., greater than 95%) and a significant potential decay (e.g., -0.78 volts), such as station K4+720. Moderate Risk Point: Select one point with a moderate wire breakage probability (e.g., 60%-80%) but a significant potential decay, such as station K7+150. Contradictory Point: Select one point with an extremely low potential (e.g., -0.75 volts) but the model initially judges a very low wire breakage probability (less than 20%), such as station K2+300. High Probability Safe Point: Select two points judged safe by the model (wire breakage probability less than 5%, normal potential -0.90 volts) as a control group, such as station K0+500 and K9+500.

[0068] Minimally invasive verification was performed at each selected verification point. At station K4+720, operators excavated a small test pit to the outer surface of the pipeline. Using a specialized miniature drilling machine equipped with a diamond thin-walled drill bit, a pilot hole with a diameter of only 12 mm and a depth of 45 mm was drilled in the concrete protective layer (just outside the steel wire layer). Subsequently, an 8 mm diameter flexible fiber optic endoscope probe was inserted into the hole. The high-definition camera (2-megapixel resolution) at the end of the probe, combined with an LED light source, clearly observed and recorded the condition of the prestressed steel wires at the bottom of the hole. The video showed that two steel wires were broken at this point. This process was recorded as the "real physical state verification result": station K4+720, two broken wires. The same method was used to verify other points, such as K2+300, where the steel wires were found to be intact, but the concrete was damp. All verification operations took approximately four hours, had minimal impact on the pipeline structure, and the drilled holes were subsequently repaired with a special polymer mortar.

[0069] Model Accuracy Evaluation and On-Site Incremental Learning: After validation, the actual results at five points were compared with the model's initial output to generate a model accuracy evaluation report. The report showed that the diagnoses of the three high-risk / contradictory points were basically correct, but the initial wire breakage probability at one safety control point (K9+500) was 3%, while actual validation revealed one broken wire (false negative). Simultaneously, at the contradictory point K2+300, the model underestimated the signal anomaly caused by concrete saturation. Due to the errors (false negatives are unacceptable), the system initiated the on-site incremental learning process. The software packaged the "complex impedance spectrum data - actual wire breakage status (number of wires)" pairs from the five validation points into an incremental training sample set containing five samples. On a mobile workstation equipped with a GPU, the underlying structure of the pre-trained wire breakage recognition model was invoked, keeping the parameters of the feature extraction layer essentially unchanged, and only unfreezing the parameters of the last three fully connected layers. Using this incremental sample set, rapid training was performed for fifty batches (approximately eight minutes) at a small learning rate of 0.001. The training objective was to minimize the mean square error between the predicted number of broken wires and the actual number of broken wires. After training is complete, save the fine-tuned model parameters.

[0070] Model reprocessing and final high-confidence correlation analysis: Using the fine-tuned and optimized new model, the complex impedance spectrum data sequence of 10,000 measurement points along the entire line was reprocessed. In the updated wire breakage probability and density distribution map, the wire breakage probability in the area near K9+500 was corrected from 3% to 65%, and the wire breakage density was corrected from 0.11 wires per meter to 1.5 wires per meter. Meanwhile, the K2+300 area was labeled as the "concrete saturation influence zone" rather than the wire breakage zone.

[0071] Using this diagnostic spectrum, validated and optimized with field data and better suited to the specific conditions of this project, a final spatial overlay and correlation analysis was performed with the actual polarization potential spectrum. The analysis results reclassified the K9+500 region as a "fracture-dominated failure zone," while the K2+300 region was classified as a "pure electrochemical failure zone caused by changes in the environmental medium." The final comprehensive evaluation report will include a complete "Model Validation and Optimization Record," listing the location of each validation point, validation results, initial model output, optimized output, and errors, and stating that "the model used in this diagnosis has undergone adaptive optimization based on five field validation points, and the average absolute error of the optimized model in identifying the number of broken wires at the validation points is 0.4." This report greatly enhances the engineering credibility of the diagnostic conclusions.

[0072] Compared with existing methods that directly apply pre-trained AI models, this implementation method solves the core bottlenecks in its application in serious engineering scenarios: the lack of "model generalization ability" and "result credibility".

[0073] This implementation method "guides verification with intelligence and optimizes intelligence with verification." It creatively transforms a small amount of high-cost direct verification into "high-value fuel" that drives and optimizes low-cost non-destructive intelligent diagnostics, forming a reliable diagnostic workflow that is self-verifying and self-improving. This not only solves the trust problem of implementing AI models in engineering, but also defines a new generation of infrastructure inspection paradigm that is human-machine collaborative and virtual-real interactive, making the results of non-destructive testing truly authoritative enough to guide critical engineering decisions.

[0074] In another technical solution, step S1 specifically includes the following sub-steps: S1a. Rapid vehicle-mounted magnetic anomaly survey: Directly above the PCCP pipeline to be tested, a testing vehicle equipped with a 16-channel high-precision fluxgate sensor array (sampling rate ≥100 Hz, resolution ≤1nT) is driven along the pipeline route at a constant speed of no less than 20 kilometers per hour; the sensor array measures the vertical and horizontal components of the surface magnetic field generated by the cathodic protection current of the pipeline in real time and synchronously, forming a surface magnetic field intensity spectrum continuously distributed along the pipeline axis; S1b, Intelligent location of abnormal protection current density sections: Based on the surface magnetic field strength map, the distribution trend of protection current density along the pipeline axis is quickly calculated and identified through the magnetic field-current inversion algorithm, and suspected abnormal sections with significant abrupt changes, attenuation or distortion of current density are located. The length range of suspected abnormal sections is 50-200 meters. S1c, Optimized deployment of distributed optical fiber sensing units: Distributed optical fiber current sensing units are precisely deployed only within each suspected abnormal section located in step S1b and within a 50-meter extension range upstream and downstream; For the parts of the pipeline that are not identified as suspected abnormal sections, distributed optical fiber current sensing units are not deployed, and only the surface magnetic field survey data obtained in step S1a is retained as background reference.

[0075] In the above technical solution, this embodiment takes a 30-kilometer-long HM water supply project PCCP pipeline as the object, and the goal is to complete a rapid preliminary assessment and key detailed investigation and location of its cathodic protection status within one working day (eight hours).

[0076] Rapid vehicle-mounted magnetic anomaly detection implementation: The detection team first used an engineering vehicle equipped with a dedicated detection system. A sensor bracket was fixedly mounted in the middle of the vehicle's chassis, on which sixteen high-precision fluxgate sensors were arranged longitudinally at 0.5-meter intervals, forming a multi-channel sensor array. The center of the array was synchronized with the vehicle's GPS and ranging wheels.

[0077] During operation, the vehicle travels at a stable speed of 25 kilometers per hour along the inspection road directly above the pipeline. While the vehicle is in motion, sixteen sensors simultaneously and continuously collect surface magnetic field data at a sampling rate of 100 times per second, recording the precise latitude and longitude coordinates and mileage markers for each data point. The measurements primarily target the weak magnetic field generated by the cathodic protection current of the underground pipeline, focusing on collecting the vertical and horizontal components of the magnetic field. Data collection for the entire 30-kilometer pipeline can be completed in approximately one and a half hours. The raw data is transmitted back to the onboard industrial control computer in real time, generating a continuous surface magnetic field strength map along the entire route. This map uses color depth to represent magnetic field strength, visually showing that in most sections, the magnetic field strength remains stable at a baseline level of approximately 500 nanoteslas.

[0078] Intelligent location of suspected abnormal sections: After data acquisition, the onboard processing software immediately runs the magnetic field-current inversion algorithm. This algorithm, based on a simplified model of the Biot-Savart law, converts continuous magnetic field data into an equivalent pipeline axial protection current density distribution trend curve. The software has an automatic diagnostic threshold: when the equivalent current density value of a certain pipeline section suddenly drops by more than 30% compared to the upstream stable value, or exhibits severe sawtooth fluctuations, it is marked as abnormal.

[0079] For example, after processing the data, the algorithm found that between chainage K12+300 and K12+700, the equivalent current density plummeted from 120 mA / m at the baseline to 70 mA / m, and remained low within this range. Between chainage K22+000 and K22+150, the current density curve exhibited violent oscillations. Based on this, the software automatically identified five suspected anomalous sections, each between 100 and 200 meters in length. The system generated a "Quick Survey Report" and highlighted these five sections on a pipeline route diagram, providing the starting and ending chainages, magnetic field anomaly characteristics, and the percentage of equivalent current attenuation for each section.

[0080] Optimized Fiber Optic Deployment and Precise Detailed Inspection: Guided by the rapid survey report, the inspection team no longer needed to deploy distributed fiber optic cables along the entire 30-kilometer pipeline. The work mode shifted to "precision deployment." The team drove their engineering vehicle directly to the first suspected anomaly section, such as K12+300 to K12+700. They precisely deployed, coupled, and debugged the distributed fiber optic current sensing units only within this section and its upstream and downstream extensions of 50 meters (i.e., K12+250 to K12+750), a total of 500 meters. After deployment, high-precision current density measurements and all subsequent analysis steps were performed in this section. After completing the detailed diagnosis of this section and retrieving the fiber optic cable, the team quickly moved to the next suspected anomaly section and repeated this "fixed-point deployment - precise measurement" process. For the remaining large pipeline sections (approximately 90% of the total length) where the report showed a stable magnetic field and no abnormal equivalent current density, no fiber optic cables were deployed; only the rapid survey data was retained as a "normal background" reference. This model compresses the fiber optic cable laying and full-line precision measurement work, which might have taken several days to complete, into a single day. It focuses on key sections that account for less than 10% of the total length, greatly improving work efficiency and minimizing the occupation and disturbance to the ground above the pipeline.

[0081] Compared to traditional vehicle-mounted potential detection systems: Existing vehicle-mounted systems are merely mobile versions of traditional point measurements, outputting a "snapshot of surface potential" that may be affected by soil unevenness. Their data has a single purpose and limited accuracy. The first-stage (magnetic measurement) output of this implementation is not a final conclusion, but rather a "smart diagnostic map" to guide subsequent actions. It uses physical inversion to transform the magnetic field into a more engineering-significant current density trend, and based on this, makes crucial decisions about "where further investment is needed," forming a complete logical chain of "problem discovery through survey - problem confirmation through detailed investigation." Differences in detection principle and information quality: Traditional vehicle-mounted systems measure potential, which is easily affected by environmental factors (such as road asphalt and soil moisture), resulting in a low signal-to-noise ratio. This implementation measures the magnetic field; the magnetic field signal is less affected by the upper medium and can more directly reflect changes in the current within the pipeline. Therefore, the reliability of the rapid survey phase is higher, providing a more credible basis for subsequent decisions.

[0082] The two-tier architecture proposed in this implementation is not simply about accelerating a single technology, but rather creatively constructing a hierarchical, decision-driven collaborative detection system. It organically integrates "wide-area rapid perception" with "local fine-grained diagnosis," using the former to ensure efficiency and the latter to guarantee depth, seamlessly connecting the two through intelligent information flow. This is not merely an improvement on a single detection technology, but a paradigm shift in long-distance infrastructure inspection processes, achieving the engineering goal of acquiring the most critical information at the optimal cost without affecting project operations.

[0083] In another technical solution, after completing the comprehensive evaluation and positioning in step S4, the solution also includes data fusion and unified spatiotemporal archive construction steps, specifically: Establish a unique digital pipeline archive associated with the PCCP pipeline under test: All source and derived data generated during this inspection, including the spatial geographic coordinates of each measurement point, acquisition timestamp, original value of axial protection current density, original data of multi-frequency impedance spectrum, surface magnetic field data, true polarization potential obtained from inversion calculation, wire breakage diagnosis results, model calibration parameters, simulation prediction data, and the aforementioned comprehensive evaluation report, are synchronously stored in the digital pipeline archive with a unified data structure; the digital pipeline archive supports multi-dimensional data association queries, comparative analysis, and visualization based on spatial location and time.

[0084] In the above technical solution, this implementation assumes that a complete inspection process for a 10-kilometer PCCP pipeline has been completed (which may include rapid general survey, detailed fiber optic inspection, model calibration, broken wire identification, etc.), and a final comprehensive evaluation report has been generated. Creation and association of digital pipeline archives: After the inspection task is completed, the data processing center initiates the archive construction process. First, the system queries the central database based on the unique identifier of the inspected pipeline (e.g., "HM Water Supply Project - North Trunk Line - Section A - PCCP Pipeline - Asset Number PC2023001"). If this is the first inspection of the pipeline, the system will automatically create its own digital pipeline archive. The core of the archive is a master record in a relational database, associated with a dedicated cloud storage directory, the directory path of which is bound to the asset number. For example, the system creates the path " / Project_HM / Asset_PC2023001 / " in cloud storage to store all historical data for this pipeline. During archive creation, basic pipeline design information is simultaneously entered, including a total length of 10 kilometers, a pipe diameter of 2.4 meters, commissioning date, drawing number, etc. Multi-source data standardization and structured storage: Next, the system executes an automated data archiving pipeline. This pipeline calls a dedicated data parser and converter to process various files generated during this inspection, converting them into a unified format and injecting them into the archive database. Spatiotemporal benchmark unification and raw data return: The system first assigns a unique "task ID" to this inspection task, such as "Survey_20231027_001". All data will be tagged with this ID. Then, the raw data from each channel is processed: For data collected by the distributed fiber optic sensing unit, its binary data stream is parsed, and the timestamp, mileage station number (accurate to meters), and current density value (unit: milliamperes per square meter) corresponding to each sampling point are extracted to form a data table, which is stored in the "Raw Current Density Table" in the database and associated with the task ID. For data files from the multi-frequency impedance spectrometer, the station number, impedance amplitude, and phase angle corresponding to ten characteristic frequencies (from 100 Hz to 100 kHz) of each measurement point are extracted and stored in the "Raw Impedance Spectrum Table". For vehicle-mounted magnetic anomaly survey data, the synchronous records of GPS trajectory and magnetic field strength (in nanotesla) are analyzed and stored in the "Original Magnetic Field Data Table". For endoscopic videos of field verification points, the video files themselves are stored in the " / Original Media / " subfolder of the cloud directory, and the video summary information (verification point station number, verification time, conclusion "two broken wires found") is stored in the "Field Verification Record Table" of the database. Structured storage of process and result data: The system then archives the key data and final results generated during the analysis process: The parameter files (such as mesh files and material property configuration files) of the calibrated final version of the current-potential field coupling inversion model are packaged and stored in the " / Model Parameters / " subdirectory of the cloud directory, and its version number and the verification point information used for calibration are recorded in the "Model Version Table" of the database.The final output of the full-line "True Polarization Potential Spectrum" data is stored as a "Potential Distribution Table" in the database at a density of one data point per meter. Each record includes the station number, potential value, and task ID. The "Wire Breakage Probability and Density Distribution Spectrum" data is stored as a "Wire Breakage Diagnosis Table," including the station number, wire breakage probability percentage, and estimated number of broken wires per meter. The core conclusions from the comprehensive assessment report—the list of risk sections—are parsed and stored in the "Risk Section Table," recording the start and end station numbers, length, average potential, and dominant failure mode (e.g., "Wire Breakage Shielding Type") for each section. Visualization, querying, and comparison functions of the archives are implemented: After archiving, the archive information is made available to authorized users on a dedicated "Pipeline Health Management Platform." The platform provides a web-based interactive interface. After logging in, users select the target pipeline "PC2023001" in the asset tree, and the main interface displays the pipeline's plan view. Users can select the current inspection task "Survey_20231027_001" in the "Inspection History" panel on the right. The map immediately renders the "True Polarization Potential Spectrum" of this detection using color gradients, and identifies risk sections are marked with flashing red boxes. Users can overlay a "Filament Density Distribution" layer as a bar chart on the pipeline. Users can perform complex spatiotemporal queries. For example, by entering the station interval "K3+000 to K3+500" in the query box and selecting the comparison task "Survey_20221015_001" (the detection of the same section last year), the platform will extract all potential and filament breakage data for that section from the archive database and generate a dual-axis trend comparison chart, clearly showing the specific changes in potential decay and filament breakage development in that section over the past year. All query results can be exported as reports in a standard format.

[0085] Compared to traditional project data archiving based on paper or PDF reports, which archives test reports as the final deliverable with fixed, unstructured images and text, this implementation method preserves all raw, structured, spatiotemporally tagged data, process data, and result data generated during the testing process, forming "data assets" that can be directly understood and processed by computers. This allows data to be continuously mined, analyzed, and reused, rather than simply being "archived." The difference in data correlation and traceability is enormous: in traditional reports, a conclusion (such as a low potential at K5+000) is difficult to trace back to its original measurement data and processing parameters. This implementation method, through a unified spatiotemporal index (station number, task ID) and relational database, fully records the lineage relationships between data. Users can easily see which original measurements a risk conclusion is based on and which version of the model was used to calculate it, achieving full-process traceability and greatly enhancing data credibility and analytical depth.

[0086] The data fusion and unified spatiotemporal archive construction steps proposed in this implementation method are a systematic solution aimed at freezing and enhancing the value of data. Through a standardized data governance framework, it transforms one-off inspection service outputs into "data capital" that can continuously generate interest throughout the pipeline's entire lifecycle of digital management. This not only solves the data integration challenges brought about by the fusion of multiple technologies but also provides a solid data foundation for predictive maintenance, big data analysis, and scientific decision-making, serving as a key guarantee for maximizing the value of the entire advanced inspection methodology system.

[0087] In another technical solution, after the digital pipeline archive is built, it also includes long-term trend analysis and early warning steps, specifically: In the digital pipeline archive, the true polarization potential and broken wire density data obtained from each detection at the same spatial location point are extracted in time sequence. Based on time series data, a potential decay trend prediction model and a wire breakage development rate prediction model were constructed respectively. When the predictive model indicates that the potential value of a specific section will drop below the protection standard threshold within a predetermined period, or that the broken wire density will exceed the safety threshold, the system automatically generates a warning alert and identifies the key time points in the risk evolution.

[0088] As a preferred method, the construction methods for the potential decay trend prediction model and the wire breakage rate prediction model are as follows: Data organization: For the measuring point at the same spatial location in the pipeline archive, extract the detection time t of each time. i (i=1, 2,..., N), where N is the number of detections and the corresponding actual polarization potential value E. i Or broken fiber density value D i Construct time series {t i E i} or {t i D i Considering that pipeline degradation typically exhibits monotonic characteristics, predictive models that include trend terms are preferred when constructing the model.

[0089] The Holt's Linear Trend Exponential Smoothing Model was selected, which includes a horizontal component l. t With trend component b t Its update equation is: l t = α × E t + (1-α) × (l t-1 + b t-1 ) b t = β× (l t-l t-1 ) + (1-β) × b t-1 Where α and β are smoothing parameters, ranging from 0 to 1, based on the historical potential sequence {E i The maximum likelihood estimation method is used to optimize the solution for α and β. The prediction formula is E. t+h = l t + h × b t Where h is the forward prediction step size (the unit can be set to months). Simultaneously, the model can calculate the prediction interval to quantify the uncertainty of the prediction results.

[0090] Prediction model for the rate of fiber breakage: Considering that the fiber breakage process may exhibit exponential or polynomial acceleration characteristics, a time series regression model containing a quadratic growth term is selected, which has the following form: D t = c + ω1× t + ω2× t 2 +ε t Where c is the intercept term, ω1 and ω2 are the coefficients of the linear and quadratic terms, respectively, and ε t For the random error term (satisfying ε) t ~N (0, s 2 ), s 2 (The variance of the error term). Based on the historical broken wire density sequence {t i D i The least squares method is used to estimate the model parameters; by substituting the future time t+h into the model, the predicted value D of the broken wire density at the corresponding time can be obtained. t+h .

[0091] Warning generation: Pre-set protection potential threshold E th (e.g., -0.85V) and the safety broken wire density threshold D th (e.g., 2.0 roots / meter). For each spatial measurement point, using its corresponding prediction model, predict the potential value E in the next H months (e.g., 24 months). pred and its upper limit of confidence interval, or the predicted value of broken filament density D pred If E pred The upper limit of the confidence interval is greater than E th , or D pred Greater than D th If this occurs, the system will automatically trigger an early warning and identify the expected time point T when the risk is predicted to occur. risk .

[0092] In the above technical solution, this embodiment takes a PCCP pipeline that has been in operation for five years and has completed three (e.g., once a year) comprehensive non-destructive rapid inspections as an example. Its asset number is PC2018001 and a complete digital archive has been established.

[0093] Historical Data Time Series Extraction and Preprocessing: Operators or system-timed tasks on the pipeline management platform initiate the trend analysis module. The module first automatically extracts key data from the pipeline's PC2018001 digital archive, covering all historical inspection tasks within a specified analysis section (e.g., the entire pipeline or a key focus section). For example, the system extracts the corresponding data for three inspections (October 2021, October 2022, and October 2023) at station K5+000: True polarization potential values: -0.915V, -0.890V, -0.868V; Estimated broken wire density values: 0.2 wires per meter, 0.5 wires per meter, 0.9 wires per meter. The system arranges this data chronologically, forming two independent time series datasets for each analysis point. Data standardization and spatial alignment are performed before data entry to ensure comparability. For points with missing data (e.g., not covered by a particular inspection), the system marks them and uses appropriate data interpolation algorithms during modeling.

[0094] Construction and Training of Trend Prediction Models: The system utilizes its built-in machine learning analysis engine to model and analyze the extracted time-series data. Potential Decay Trend Prediction Model: For potential data, the system uses an exponentially smoothed state-space model for fitting. Taking point K5+000 as an example, the model analysis reveals that the potential at this point decays linearly at an average rate of approximately -0.023 volts per year. The model calculates the confidence interval for this decay trend. Based on this, the model extrapolates and predicts potential values ​​for future periods. For example, it predicts that the potential at this point will decrease to -0.835 volts in eighteen months (i.e., April 2025), with a 90% probability of falling between -0.825 volts and -0.845 volts. Wire Breakage Development Rate Prediction Model: For wire breakage density data, considering its typical "start-up-acceleration" characteristic, the system uses a time-series model with a growth term or a specific regression model. The model analyzes the data at point K5+000 and determines that its wire breakage development may have entered the initial acceleration stage, predicting that its wire breakage density may reach 1.8 wires per meter in twenty-four months.

[0095] Early Warning Generation, Issuance, and Decision Support: After the predictive model runs, the system automatically compares the prediction results with the preset engineering safety thresholds. Assuming the effective potential threshold for cathodic protection of the pipeline is -0.850 volts and the broken wire density alarm threshold is 2.0 wires per meter, the system detects that the predicted potential value (-0.835 volts) at point K5+000 will be lower than the protection threshold in eighteen months, with a high probability. Simultaneously, multiple points in the adjacent section from K4+900 to K5+100 show a similar trend. Based on this, the system automatically generates a Level 1 early warning message. The warning message clearly states: Warning object: Pipeline PC2018001, section from chainage K4+900 to K5+100 (200 meters in length). Warning content: Based on the trend analysis of data from the past three years, the actual polarization potential of the steel cylinder in this section shows a continuous decreasing trend, and it is predicted that it will generally be lower than the effective protection standard in approximately eighteen months (expected April 2025), posing a risk of accelerated corrosion. Recommended measures: It is recommended that this section be included as a key area for review in the next inspection (e.g., six months later), and preventative maintenance measures should be planned in advance. This early warning information is automatically sent to designated pipeline engineers, maintenance managers, and other responsible personnel via the management platform's message center and email. The platform interface marks this section as a "Trend Warning Zone" on the map and displays it in a flashing amber color. The early warning information is accompanied by a detailed analysis report, including historical data curves, forecast curves, confidence intervals, and key time points.

[0096] Compared to traditional methods based on fixed-cycle (e.g., annual) detection and threshold comparison, which merely determine "whether the current standard is met" after each detection (similar to marking a "qualified / unqualified" point on a timeline), this implementation method analyzes time-series vectors to not only know the current state but also determine its direction and speed of change, predicting future states. This is equivalent to transforming from looking at a "photo" to watching a "dynamic video and predicting the next few frames," resulting in a qualitative improvement in information dimensionality. From "uniform cycle" to "state-based dynamic cycle": the detection cycle of the traditional method is fixed and arbitrary. Based on trend prediction results, this implementation method can recommend differentiated and optimal next detection or intervention times for pipeline segments in different states. For stable segments, the detection interval can be extended; for rapidly deteriorating segments, the interval can be shortened and early warnings provided, achieving precise and efficient allocation of maintenance resources.

[0097] This implementation method, by constructing a predictive analysis engine based on historical archive data, achieves for the first time in the field of PCCP pipeline cathodic protection management the quantitative prediction and forward-looking early warning of asset performance degradation trajectory. It marks a leap in pipeline operation and maintenance from a passive medical model relying on periodic "check-ups" to a proactive health management model based on continuous "health monitoring" and "disease prediction." It is a core manifestation of intelligent and refined asset management, possessing significant management innovation and economic value.

[0098] In another technical solution, for the underprotected risk section identified in step S4, a failure root cause contribution quantification analysis step is also included. Specifically, for the underprotected risk section, a multi-scenario simulation comparison model is constructed based on the current-potential field coupling inversion model. The multi-scenario simulation comparison model simulates and calculates the protection potential distribution under the following conditions: only the current diagnostic wire breakage state exists, only the measured environmental medium parameters exist, and other single assumed failure factors exist. By comparing the difference in the degree of agreement between each simulation result and the measured potential distribution obtained from the inversion in step S3, the independent contribution and coupling contribution of the wire breakage factor, environmental medium factor, and other factors to the current potential decay result are quantified and evaluated, and a root cause quantification analysis report is generated.

[0099] In the above technical solution, this implementation takes a pipeline section identified as having a "composite failure" as an example. This section is located between chainage K8+100 and K8+300, with a length of 200 meters. Preliminary testing showed that the average true polarization potential of this section was -0.78 volts, while the broken wire identification model showed a broken wire density of three wires per meter, and soil resistivity survey data showed that the soil in this area was relatively dry. Construction and setting of a multi-scenario simulation comparison model: After completing detailed testing and calibration modeling of this section, the engineer started the root cause analysis module in the digital twin model platform. The system first copied a high-precision current-potential field coupling inversion model that had been calibrated with field data and accurately reflected the current comprehensive state of the section as the "baseline model". Subsequently, engineers defined three independent scenario simulation models in this module: Scenario 1 (Wire Breakage Only): On a copy of the baseline model, only the currently diagnosed wire breakage state is retained (i.e., the conductivity of the prestressed steel wire is set to the broken state at the corresponding location in the model), but the soil resistivity parameter in the model is restored to the typical value (e.g., 50 ohm-meters) for most areas along the pipeline, and the concrete moisture parameter is set to the standard value. This scenario simulates "how much impact the current wire breakage defect alone will have under normal environmental conditions." Scenario 2 (Environmental Factors Only): On another copy of the baseline model, the steel wires are kept continuous and intact without any wire breakage defects, but the soil resistivity parameter for this section in the model is set to a high value measured in the field (e.g., 120 ohm-meters), and the equivalent resistivity of the concrete may be adjusted based on impedance spectrum data. This scenario simulates "how much the protective effect will decrease under the current harsh environmental conditions, even if the structure is intact." Scenario 3 (Comprehensive Baseline Scenario): This is the model that reflects the actual state, including both the current wire breakage state and the high soil resistivity environment. Simulation Calculation and Contribution Measurement Analysis: After defining the scenarios, the system automatically runs simulation calculations for the three models, obtaining the steel cylinder potential distribution curves for the 200-meter section under the three assumptions. The analysis software extracts key data for comparison. For example, taking the data at the midpoint K8+200 of this section: Measured / Comprehensive Scenario 3 Potential: -0.78V (this is the measured inversion value, used as a comparison benchmark). Scenario 1 (Wire Breakage Only) Simulated Potential: -0.84V. Scenario 2 (Environment Only) Simulated Potential: -0.87V. Ideal Intact State (No Wire Breakage, Standard Environment) Theoretical Potential: -0.92V (used as the zero-point reference for analysis). The system starts the quantitative evaluation algorithm to calculate the independent contribution and coupled contribution of each factor: Independent Contribution Calculation: Wire Breakage Independent Contribution = |Ideal Potential - Scenario 1 Potential| = |-0.92V - -0.84V| = 0.08V attenuation. Environmental independent contribution = |Ideal potential - Scenario 2 potential| = |-0.92V --0.87V| = 0.05V attenuation.Coupling contribution calculation: Theoretical superposition value = Ideal potential - (Independent contribution of broken wire + Independent contribution of environment) = -0.92V - 0.13V = -0.79V. Measured comprehensive attenuation value = Ideal potential - Measured potential = -0.92V - -0.78V = 0.14V. Coupling contribution = Measured comprehensive attenuation value - (Independent contribution of broken wire + Independent contribution of environment) = 0.14V - 0.13V = 0.01V. This positive value indicates that in this analysis, there is a weak synergistic deterioration effect (coupling effect) between broken wire and adverse environmental conditions, resulting in a final potential that is 0.01V worse than the simple sum of their independent effects.

[0100] Contribution ratio allocation: Based on the above values, the system calculates that in the current comprehensive potential decay (0.14 volts): the contribution ratio of wire breakage factors is approximately 57%. The contribution ratio of environmental factors is approximately 36%. The contribution ratio of the coupling effect of the two is approximately 7%.

[0101] Root Cause Analysis Report Generation: The system automatically generates a "Quantitative Analysis Report on the Root Causes of Failure." The report clearly presents a comparison of simulated potential curves under three scenarios in a graphical format. The core conclusions of the report clearly state: Primary Contradiction: In this complex failure section (K8+100 to K8+300), the main cause of the substandard potential is the breakage of the prestressed steel wire, contributing more than 50%. Secondary Contradiction: Dry soil and high resistivity are significant aggravating factors. Effect Analysis: There is a slight negative synergistic effect between the two. Maintenance Decision Recommendations: Based on the quantitative analysis, it is recommended to prioritize structural repair or local enhanced cathodic protection measures to overcome the wire breakage shielding effect, while evaluating the cost-effectiveness of local resistivity reduction treatment (such as replacing backfill soil) in high soil resistivity sections. The report quantifies that if only the broken wire is repaired (assuming repairability), the potential can be expected to increase to approximately -0.84 volts; if only the environment is improved, the potential can only be increased to approximately -0.87 volts, neither of which can fully meet the standard, thus strongly demonstrating the necessity and priority of comprehensive measures.

[0102] Compared to comprehensive judgments relying on expert experience, this implementation method shifts from "subjective weighting" to "objective quantification." Traditionally, when faced with the coexistence of broken fibers and poor soil quality, experts rely on observations to "roughly judge" which factor is more significant, often concluding that "the problem is likely mainly due to broken fibers" or "the soil and environment also have a large impact," which is highly subjective and ambiguous. This implementation method, through simulation experiments with controlled variables, directly translates the influence of each factor into numerical values ​​and specific percentages of potential differences, providing irrefutable objective data support and transforming decision-making from "I feel" to "data shows." It also reveals the hidden dimension of "coupling effects": traditional analysis is almost incapable of recognizing and assessing the interactions between multiple failure factors. This implementation method, by comparing the sum of independent effects with the combined effect, quantifies and reveals the "coupling contribution" for the first time. This allows engineers to realize that in some cases, addressing the two problems separately may not achieve the desired results, thus enabling more foresight in designing comprehensive maintenance solutions.

[0103] In another technical solution, in step S1b, the magnetic field-current inversion algorithm also integrates known geological survey data and historical excavation records along the pipeline route; The intelligent positioning process uses the fused data to filter out environmental interference and correct geological stratification in the magnetic field intensity map, and weights the credibility of suspected abnormal sections located by the algorithm based on the spatial location of known defects or interference points in historical records. Output a list of suspected abnormal sections with different confidence levels and corresponding decision recommendations. The decision recommendations include conducting a detailed fiber optic survey immediately for high-confidence sections, suggesting supplementing medium-confidence sections with close-range magnetic surveys or ground-penetrating radar scans, and marking low-confidence sections as long-term monitoring points of concern.

[0104] In the above technical solution, this embodiment takes a 20-kilometer PCCP pipeline located in a complex suburban area as an example. The area is known to contain underground cables, ancient riverbeds, and some backfill soil layers.

[0105] Preparation and Integration of Multi-Source Prior Data: Before conducting a rapid survey of vehicle-mounted magnetic anomalies, the inspection team first imported and associated multi-source prior data for this section of the pipeline into the system: Geological Survey Data: Detailed geotechnical engineering survey reports for this section were obtained from the engineering archives, and soil resistivity stratification data were extracted. For example, the report showed that in the section from chainage K10+000 to K11+000, there was a gravel and gravel ancient riverbed sedimentary layer with a thickness of about 0.5 meters and a resistivity as high as 150 ohm-meters two meters below the surface. Historical Excavation and Maintenance Records: All excavation records for this pipeline over the past ten years were extracted from the maintenance database. For example, the records showed that at chainage K15+200, the anti-corrosion layer was repaired due to damage caused by third-party construction; at K5+500, a suspected anomaly was excavated two years ago, but no problems were found with the pipeline itself, and it was determined to be interference from underground abandoned metal components. Known Interference Source Register: Accurate routing diagrams of known parallel or crossing power cables (10 kV, approximately 5 meters away from the pipeline) along the route were collected and entered. Intelligent Data Processing and Anomaly Location: After the detection vehicle completes magnetic field data acquisition, the processing software runs an enhanced magnetic field-current inversion algorithm. This algorithm operates in the following steps: Environmental Interference Filtering: The algorithm first identifies and filters out regular background magnetic fields generated by known fixed interference sources (such as parallel power cables). For example, by comparing the location of the power cable with the magnetic field spectrum, the algorithm identifies a stable 50 nanotesla background field in the K8+000 to K8+500 section and subtracts it from the original data. Geological Stratigraphic Correction: Based on geological survey data, the algorithm corrects the magnetic permeability and conductivity of different strata. For example, in the paleochannel section (K10+000 to K11+000), due to the presence of a high-resistivity gravel layer, the magnetic field signal generated by the protective current will experience abnormal attenuation when propagating to the surface. The algorithm will perform "signal recovery" correction on the magnetic field measurements in this section based on the stratum model, making it comparable to the data from adjacent normal soil areas. Historical Information Weighting: The algorithm spatially matches the initially calculated current density trend anomalies with historical records. For areas historically verified as false signals (such as K5+500), even if the current magnetic field test shows a slight anomaly, the algorithm will automatically lower its confidence weight by 50%. For areas that have experienced real damage (such as K15+200), even if the current anomalous signal is not strong, their historical record will increase their attention and their weight will be increased by 30%.

[0106] Output of tiered decision-making recommendations: After processing, the system does not simply output an anomaly point map, but generates a structured "Intelligent Survey Diagnostic Report." The report categorizes suspected anomaly sections into three confidence levels and provides targeted recommendations: High-confidence sections (confidence > 80%): For example, in the section from K12+300 to K12+450, the magnetic anomaly signal remains very prominent after correction (current density attenuation reaches 40%), and there are no known interference sources or historical false signal records in this area. The system marks this as a "high-confidence anomaly," and the decision recommendation is: "Immediately include this section in the priority fiber optic detailed survey targets of this detection, and recommend arranging it as the first detailed survey point." Medium-confidence sections (confidence 50%-80%): For example, in the ancient riverbed section (K10+000 to K11+000), several discontinuous weak anomalies are still found after algorithm correction. The system judges that these anomalies may be caused by residual geological heterogeneity or early, minor pipeline problems. The decision recommendation is: "It is recommended to use higher resolution, close-range (e.g., one point every five meters) magnetic resurveys at this stage, or supplement it with ground-penetrating radar scanning, to further clarify the nature of the anomaly before deciding whether to conduct a detailed fiber optic survey." Low-confidence areas (confidence <50%): For example, near K5+500 (historical false signal area) and some areas near large substations, although the raw data fluctuates, after interference filtering and historical weighting, the anomaly confidence is very low. The system marks these as "low confidence / potential interference." The decision recommendation is: "It is not recommended to invest resources in a detailed survey this time. Mark these as long-term monitoring points and continuously observe their signal change trends in subsequent rapid surveys."

[0107] The report ultimately outputs a clear roadmap to guide the inspection team in accurately and efficiently investing limited time and expensive distributed fiber optic resources in the pipeline sections most likely to have real problems.

[0108] Compared to magnetic survey interpretation techniques that rely solely on the assumption of a "clean" environment, traditional magnetic survey interpretation typically assumes that the pipeline is located in a uniform, undisturbed, and infinitely large semi-space. This inevitably introduces a large number of misjudgments in complex realities. This implementation method constructs a realistic, non-uniform "interpretation background field" by actively fusing prior geological and environmental data. It can not only identify anomalies but also distinguish whether the anomalies originate from problems within the pipeline itself or from "illusions" caused by the external environment (such as special strata or nearby cables), greatly improving the physical authenticity and reliability of signal interpretation. The "memory" and "experience-learning" capabilities of the diagnosis: Traditional methods start "from scratch" with each test, unable to utilize historical experience. This implementation method, by fusing historical excavation and verification records, endows the diagnostic system with "memory." It can learn from past successes and errors (such as verified false signal points), avoiding repeated mistakes in the same place or maintaining reasonable vigilance in areas where problems have occurred. This continuously accumulating knowledge base allows the overall judgment capability of the detection system to evolve over time.

[0109] Although the technical solutions of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and embodiments shown and described herein.

Claims

1. A non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects, characterized in that, Includes the following steps: S1. Distributed Axial Protection Current Density Measurement: Distributed fiber optic current sensing units are axially deployed along the outer wall surface or under the soil layer of the PCCP pipeline to be tested. Based on the Faraday magneto-optical effect, the distributed fiber optic current sensing units perform non-contact, uninterrupted continuous sensing of the axial circumferential magnetic field generated by the cathodic protection current of the steel cylinder inside the PCCP pipeline, thereby acquiring and recording the time-related axial protection current density data sequence continuously distributed along the pipeline axis in real time. S2. Construct a PCCP pipeline current-potential field coupled inversion model: Based on the actual structural parameters, material property parameters and environmental parameters of the PCCP pipeline, construct a multi-layer physical field finite element model including the steel cylinder, prestressed steel wire, concrete protective layer, anti-corrosion layer and soil medium; use the axial protection current density data sequence measured in step S1 as the dynamic boundary condition input of the current-potential field coupled inversion model. S3. Distributed inversion calculation of the true polarization potential of the steel cylinder: Run the current-potential field coupled inversion model, solve the coupled equations of the current field and potential field under dynamic boundary conditions, calculate the true polarization potential inside the concrete protective layer and on the outer surface of the steel cylinder in real time, and generate a true polarization potential spectrum continuously distributed along the pipeline axis. S4. Comprehensive evaluation and positioning of cathodic protection effect: The true polarization potential spectrum obtained by inversion in step S3 is compared point by point with the preset standard range of effective cathodic protection potential. Identify and locate pipeline sections whose potential values ​​fall outside the standard range, thus determining them as risk sections with cathodic protection failure or underprotection, and output a comprehensive assessment report including the geographical location and risk level of the risk section.

2. The non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects as described in claim 1, characterized in that, Between steps S2 and S3, there is also an online model calibration step, specifically: At least two potential verification points with known spatial locations are set up along the PCCP pipeline to be tested. At each potential verification point, the actual polarization potential of the outer surface of the steel cylinder at that point is directly measured by a pre-embedded or minimally invasively installed embedded reference electrode. The measured true polarization potential at each potential verification point is compared one by one with the calculated potential value at the corresponding position obtained by inversion from the current current density data based on the initial model parameters. Based on the deviation generated by the comparison, the dielectric electrical parameters of the region adjacent to the potential verification point in the current-potential field coupled inversion model are dynamically corrected by the optimization algorithm until the error between the calculated potential value and the measured true polarization potential is less than the set threshold, thereby obtaining a calibrated high-precision coupled inversion model. In step S3, the calibrated high-precision coupled inversion model is used to perform distributed inversion calculation of the true polarization potential of the steel cylinder.

3. The non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects as described in claim 2, characterized in that, Following step S4, the following steps are also included: S5. Digital twin model construction: Based on the current-potential field coupling inversion model constructed in step S2 and optimized by the online model calibration step between steps S2 and S3, and combined with the current state data collected in step S1 and the evaluation report output in step S4, a benchmark digital twin model consistent with the current actual pipeline protection state is established. S6. Virtual Intervention and Effect Simulation: In the operation interface of the benchmark digital twin model, at least one proposed cathodic protection system intervention measure is defined parametrically. The intervention measures include adding auxiliary anodes, adjusting the output parameters of the potentiostat, and locally repairing the anti-corrosion layer. Based on the definition of the intervention measures, the digital twin model updates the model parameters and boundary conditions in real time, and calculates and simulates the predicted true polarization potential distribution along the entire steel cylinder of the pipeline after the intervention measures are applied. S7. Scheme Comparison and Optimization: Compare and analyze the predicted real polarization potential distribution diagrams under each intervention scheme obtained from simulation calculations, evaluate the effect of each scheme on eliminating the original risk section, and at the same time evaluate whether it will cause new underprotected or overprotected risk areas throughout the entire line. Based on the predetermined technical and economic optimization objectives, a recommended intervention scheme is selected from the simulated schemes, and a simulation effect report and key construction parameters of the scheme are output.

4. The non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects as described in claim 1, characterized in that, Step S1 also includes: while performing distributed axial protection current density measurement, simultaneously performing non-contact multi-frequency impedance spectrum measurement along the pipeline axis to obtain complex impedance spectrum data sequences at each measurement point on the outer wall surface of the pipeline. In step S4, the comprehensive evaluation and localization also includes: inputting the complex impedance spectrum data sequence into the pre-trained wire breakage identification model to obtain the probability and density distribution map of prestressed steel wire breakage along the pipe axis; Spatial overlay and correlation analysis were performed on the probability and density distribution of broken wires and the actual polarization potential spectrum. Based on the correlation analysis results, the following failure mode types were identified and distinguished: the underprotected risk section caused by the current shielding effect due to the breakage of the prestressed steel wire, and the underprotected risk section caused by insufficient output of the cathodic protection system.

5. The non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects as described in claim 4, characterized in that, In step S4, after obtaining the broken wire probability and density distribution map and before performing spatial overlay and correlation analysis, a model field validation and adaptive optimization step is also included, specifically: Based on the probability and density distribution of broken wires and the actual polarization potential spectrum, at least two representative field verification points were selected. For each on-site verification point, the actual physical state verification results of the prestressed steel wire at that point are obtained through minimally invasive testing or internal video inspection. The actual physical state verification results are compared with the initial diagnostic results output by the broken wire identification model for that point to generate a model accuracy evaluation report. If the comparison error exceeds the preset threshold, the verification results of the on-site verification points and their corresponding real physical states will be combined with the original complex impedance spectrum data to form an incremental training sample set, and the pre-trained broken wire identification model will be subjected to on-site incremental learning and parameter fine-tuning. Using the finely tuned and optimized broken wire identification model, the complex impedance spectrum data sequence is reprocessed to generate an updated broken wire probability and density distribution map that is more suitable for the specific conditions of the current pipeline under test, which is then used for subsequent spatial overlay and correlation analysis.

6. The non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects as described in claim 1, characterized in that, Step S1 specifically includes the following sub-steps: S1a. Rapid vehicle-mounted magnetic anomaly survey: On the ground directly above the PCCP pipeline to be tested, drive a testing vehicle equipped with a multi-channel high-precision fluxgate sensor array, and drive at a constant speed of no less than 20 kilometers per hour along the pipeline route; the sensor array measures the vertical and horizontal components of the surface magnetic field generated by the cathodic protection current of the pipeline in real time and synchronously, forming a surface magnetic field intensity spectrum continuously distributed along the pipeline axis. S1b, Intelligent location of abnormal protection current density sections: Based on the surface magnetic field strength map, the distribution trend of protection current density along the pipeline axis is quickly calculated and identified through the magnetic field-current inversion algorithm, and suspected abnormal sections with significant abrupt changes, attenuation or distortion of current density are located. The length range of suspected abnormal sections is 50-200 meters. S1c, Optimized deployment of distributed optical fiber sensing units: Distributed optical fiber current sensing units are precisely deployed only within each suspected abnormal section located in step S1b and within a 50-meter extension range upstream and downstream; For the parts of the pipeline that are not identified as suspected abnormal sections, distributed optical fiber current sensing units are not deployed, and only the surface magnetic field survey data obtained in step S1a is retained as background reference.

7. The non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects as described in claim 1, characterized in that, After completing the comprehensive assessment and positioning in step S4, the process also includes data fusion and unified spatiotemporal archive construction, specifically: Establish a unique digital pipeline archive associated with the PCCP pipeline under test: All source data and derived data generated in this test, including the spatial geographic coordinates of each measurement point, acquisition timestamp, original value of axial protection current density, original data of multi-frequency impedance spectrum, surface magnetic field data, true polarization potential obtained by inversion calculation, wire breakage diagnosis results, model calibration parameters, simulation prediction data, and the comprehensive evaluation report, are all synchronously stored in the digital pipeline archive with a unified data structure. Digital pipeline archives support multi-dimensional data association queries, comparative analysis, and visualization based on spatial location and time.

8. The non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects as described in claim 7, characterized in that, After the digital pipeline archive is built, long-term trend analysis and early warning steps are also included, specifically: In the digital pipeline archive, the true polarization potential and broken wire density data obtained from each detection at the same spatial location point are extracted in time sequence. Based on time series data, a potential decay trend prediction model and a wire breakage development rate prediction model were constructed respectively. When the predictive model indicates that the potential value of a specific section will drop below the protection standard threshold within a predetermined period, or that the broken wire density will exceed the safety threshold, the system automatically generates a warning alert and identifies the key time points in the risk evolution.

9. The non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects as described in claim 4, characterized in that, For the underprotected risk sections identified in step S4, a failure root cause contribution quantification analysis step is also included, specifically: For the underprotected risk section, a multi-scenario simulation comparison model is constructed based on the current-potential field coupling inversion model; The multi-scenario simulation comparison model simulates and calculates the protection potential distribution under the following conditions: only the current diagnostic wire breakage state exists; only the measured environmental medium parameters exist; and other single assumed failure factors exist. By comparing the differences in the degree of agreement between each simulation result and the measured potential distribution obtained from the inversion in step S3, the independent contribution and coupling contribution of wire breakage factors, environmental medium factors and other factors to the current potential decay result are quantitatively evaluated, and a root cause quantitative analysis report is generated.

10. The non-destructive rapid testing method for the cathodic protection effect of PCCP pipelines in HM water supply projects as described in claim 6, characterized in that, In step S1b, the magnetic field-current inversion algorithm also integrates known geological survey data and historical excavation records along the pipeline route; The intelligent positioning process uses the fused data to filter out environmental interference and correct geological stratification in the magnetic field intensity map, and weights the credibility of suspected abnormal sections located by the algorithm based on the spatial location of known defects or interference points in historical records. Output a list of suspected abnormal sections with different confidence levels and corresponding decision recommendations. The decision recommendations include conducting a detailed fiber optic survey immediately for high-confidence sections, suggesting supplementing medium-confidence sections with close-range magnetic surveys or ground-penetrating radar scans, and marking low-confidence sections as long-term monitoring points of concern.