A method and system for detecting corrosion resistance of a cable protection pipe

Abstract

The application provides a corrosion resistance detection method and system for a cable protection pipe. The method belongs to the technical field of power engineering material detection and underground pipe network state evaluation. The method comprises the following steps: constructing a complete soil electrolyte three-electrode system; inputting a periodic stray current signal with a preset amplitude into the three-electrode system, synchronously applying a four-point bending constant load at a preset span position of the pipe body, completing multi-physical field collaborative loading state initialization, and generating an initial coupled loading detection sample; through the collaborative loading of the stray current and the four-point bending constant load, the real corrosion degradation path of the buried cable protection pipe is completely reproduced, and the deviation between the laboratory detection result and the actual field working condition is greatly reduced.

Description

Technical Field

[0001] This invention proposes a method and system for testing the corrosion resistance of cable protection pipes, belonging to the technical field of power engineering material testing and underground pipeline network condition assessment. Background Technology

[0002] Currently, the scale of underground cable laying projects for metropolitan area networks in China continues to expand, and the service life of underground cable protection pipes is generally required to reach more than 30 years. However, the current testing system has long used the traditional methods of salt spray testing and static acid and alkali immersion, which is significantly mismatched with the actual underground working conditions.

[0003] The salt spray test in the existing standard only simulates the marine atmospheric environment and cannot reproduce the fluctuations in soil resistivity under buried scenarios. and The actual corrosion conditions, including changes in ion activity, sulfate-reducing bacteria metabolism, and stray current interference in subway systems, can cause test results to deviate from the actual degradation rate on-site by several times. Conventional acid-alkali static immersion tests are only conducted on sheet-like material samples, completely ignoring pipe-specific failure modes such as gaps in the spigot and socket sealing rings, galvanic corrosion of the half-sleeve bolts, and mechanical damage to the coating at the pipe support points, thus failing to reflect the true corrosion resistance of the entire pipe.

[0004] Meanwhile, the current testing system completely lacks the evaluation dimension for stress corrosion cracking, failing to consider the three-point and four-point bending states caused by uneven soil settlement, as well as the alkaline seepage caused by concrete encapsulation. Enriched environments leave the stress corrosion sensitive window of CPVC and MPP pipes completely undetectable. Current testing methods only output binary results indicating pass or fail, failing to provide a reference value for the remaining service life of the pipe. This directly leads to an imbalance in pipe replacement decisions during maintenance, resulting in either excessive replacement leading to resource waste or delayed replacement causing cable damage. Currently, the EIS and CIPS combined positioning method used in long-distance oil and gas pipelines cannot directly adapt to the special testing requirements of small-diameter cable protection pipes with mixed materials and socket joints. The industry still lacks a targeted accelerated detection solution using a dual-field coupling of stray current and bending stress. Summary of the Invention

[0005] This invention provides a method and system for testing the corrosion resistance of cable protection pipes, in order to solve the problems mentioned in the background art above: This invention proposes a method for testing the corrosion resistance of cable protection pipes, the method comprising: S1. Construct a complete three-electrode system for soil electrolyte; introduce a periodic stray current signal of preset amplitude into the three-electrode system, and simultaneously apply a four-point bending constant load at a preset span position in the pipe body to complete the initialization of the multi-physics field collaborative loading state and generate an initial coupled loading detection sample. S2. Apply a broadband AC excitation signal to the initial coupled loading test sample to complete the full-band electrochemical impedance spectroscopy scan and collect the raw data of the full-band impedance response; perform equivalent circuit fitting calculation on the raw data of the full-band impedance response to generate an impedance characteristic parameter dataset. S3. Continuously collect electrochemical noise from the impedance characteristic parameter dataset, simultaneously extract the potential and current fluctuation sequences within a fixed time period, perform multi-level wavelet packet decomposition on the fluctuation sequences, extract the energy distribution ratio of each frequency band, normalize the energy distribution ratio of each frequency band, calculate the wavelet packet energy entropy value corresponding to the microcrack initiation stage, and generate a corrosion precursor feature dataset. S4. Perform node traversal operation on the corrosion precursor feature dataset, combine the spatial coordinate information of the distributed reference electrode array, compare the impedance characteristic parameters of each acquisition node with the wavelet packet energy entropy value, locate the corrosion hotspot areas in three special locations of the pipe body: the socket, the half sleeve, and the pipe support point, and generate the corrosion hotspot distribution sequence of the whole pipe body. S5. Multi-dimensional feature splicing is performed on the corrosion hotspot distribution sequence of the entire pipe body. The impedance feature parameters, corrosion precursor features and hotspot distribution coordinates are input into the pre-trained Bayesian multi-source fusion model to generate the corrosion degradation evolution curve of the entire pipe body. S6. Perform time-series extrapolation calculations on the corrosion degradation evolution curve of the entire pipe body, output the predicted value of the remaining service life of the cable protection pipe under test and the spatial coordinates of the weakest section, and generate a complete dataset of corrosion resistance test results.

[0006] This invention proposes a corrosion resistance testing system for cable protection pipes, the system comprising: One or more processors; Memory, used to store one or more programs; Wherein, when the one or more programs are executed by the one or more processors, the one or more processors are made to implement the method described in any one of the above.

[0007] The beneficial effects of this invention are as follows: By synergistically applying stray current and a four-point bending constant load, the actual corrosion degradation path of buried cable protection pipes is completely reproduced, significantly reducing the deviation between laboratory test results and actual field conditions. Through the joint acquisition of broadband electrochemical impedance spectroscopy and electrochemical noise, early signals of microcrack initiation in the pipe body are directly captured, allowing for early risk identification without waiting for the sample to corrode to a point where it can be weighed. Through full-domain scanning of a distributed reference electrode array, corrosion hotspots at stress concentration points such as sockets and spigots are accurately located, eliminating the need for manual disassembly and inspection of weak points in the pipe body. The output of the Bayesian multi-source fusion model directly provides the remaining service life of the pipe body and the coordinates of the weakest section, enabling both corrosion resistance comparison of new products and sampling and grading of in-service pipes. This effectively avoids material waste caused by excessive replacement during maintenance and prevents cable breakage accidents caused by delayed replacement. Attached Figure Description

[0008] Figure 1 This is a diagram illustrating the steps of the method described in this invention; Figure 2 This is a detailed flowchart of step S5 in this invention. Detailed Implementation

[0009] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0010] One embodiment of the present invention, such as Figure 1 As shown, a method for testing the corrosion resistance of a cable protection pipe includes: S1. The cable protection pipe to be tested is placed into a soil corrosion medium tank with a preset composition. A distributed reference electrode array is arranged along the axial direction on the outer wall of the pipe. The metal conductive layer of the pipe is used as the working electrode and connected to the electrochemical workstation to construct a complete three-electrode system of soil electrolyte. A periodic stray current signal of preset amplitude is passed through the three-electrode system. At the same time, a four-point bending constant load is applied at a preset span position of the pipe to complete the initialization of the multi-physics field collaborative loading state and generate an initial coupled loading test sample. S2. Apply a broadband AC excitation signal to the initial coupled loading detection sample. A full-band electrochemical impedance spectroscopy scan was completed within the frequency range to collect raw impedance response data across the entire frequency band. Equivalent circuit fitting calculations were performed on the raw impedance response data across the entire frequency band to extract three core evolution parameters: coating porosity, interface exfoliation degree, and double-layer capacitance, generating an impedance characteristic parameter dataset. S3. Continuously collect electrochemical noise from the impedance characteristic parameter dataset, simultaneously extract the potential and current fluctuation sequences within a fixed time period, perform multi-level wavelet packet decomposition on the fluctuation sequences, extract the energy distribution ratio of each frequency band, normalize the energy distribution ratio of each frequency band, calculate the wavelet packet energy entropy value corresponding to the microcrack initiation stage, and generate a corrosion precursor feature dataset. S4. Perform node traversal operation on the corrosion precursor feature dataset, combine the spatial coordinate information of the distributed reference electrode array, compare the impedance characteristic parameters of each acquisition node with the wavelet packet energy entropy value, locate the corrosion hotspot areas in three special locations of the pipe body: the socket, the half sleeve, and the pipe support point, and generate the corrosion hotspot distribution sequence of the whole pipe body. S5. Multi-dimensional feature splicing is performed on the corrosion hotspot distribution sequence of the entire pipe body. The impedance feature parameters, corrosion precursor features and hotspot distribution coordinates are input into the pre-trained Bayesian multi-source fusion model to complete the probability weighting operation of the multi-source data and generate the corrosion degradation evolution curve of the entire pipe body. S6. Perform time-series extrapolation calculations on the corrosion degradation evolution curve of the entire pipe body, output the predicted value of the remaining service life of the cable protection pipe under test and the spatial coordinates of the weakest section, and generate a complete dataset of corrosion resistance test results.

[0011] The working principle and effects of the above technical solution are as follows: By employing a multi-physics field collaborative loading detection method, the complex corrosion and stress environment of the cable protection pipe during actual service can be recreated, improving the accuracy and realism of the pipe's corrosion resistance performance testing. Through the combined application of full-band electrochemical impedance spectroscopy and wavelet packet signal decomposition, early corrosion characteristics of microcrack initiation on the pipe surface can be accurately captured, effectively improving the accuracy of identifying early corrosion risks. Corrosion evolution analysis through multi-source data fusion can refine the corrosion differences at different locations within the pipe body, accurately locate various structurally weak areas, and reduce detection bias caused by single detection parameters. Through a time-series extrapolation analysis mode, the remaining service life and weakest failure location of the pipe can be accurately predicted, avoiding the problem that conventional testing can only determine the instantaneous corrosion state and cannot predict long-term degradation trends. This ensures both the comprehensiveness of the test results and improves the reliability of cable protection pipe corrosion detection and life assessment.

[0012] In one embodiment of the present invention, S1 includes: S11. Prepare multi-component composite soil corrosion medium, control the three basic physicochemical indicators of water content, salt content, acidity and alkalinity in the medium to a fixed threshold range, and uniformly fill the prepared corrosion medium into the corrosion-resistant sealed medium tank to maintain the overall static immersion time of the medium to reach the standard static immersion period. S12. Clean all floating dust, oil stains, and oxidation attachments from the outer wall of the cable protection pipe to be tested. Mark continuous sampling points at equal intervals along the longitudinal direction of the pipe. Paste insulating pads on the surface of each point. Fix a distributed reference electrode array on the surface of the pads to ensure that all electrode probes are tightly fitted to the outer wall of the pipe without gaps. S13. Strip the protective coating from the end of the cable protection pipe to expose the continuous internal metal conductive layer. Use a low-impedance conductive clamp to hold the exposed metal area. Connect the clamp to the working electrode port of the electrochemical workstation through a shielded wire. Bury a saturated calomel counter electrode at the bottom of the medium tank. Combine the completed reference electrode array to build a three-electrode system for soil electrolyte. S14. Retrieve the built-in signal output module of the electrochemical workstation, set three types of output parameters: fixed amplitude of periodic stray current, fluctuation period, and offset coefficient, continuously supply a continuous current signal to the three-electrode system, and simultaneously start the four-point bending loading device. Set a constant bending load value according to the nominal outer diameter of the pipe, and apply the load evenly to the two preset span areas of the pipe body. S15. Maintain the synchronous operation of current signal transmission and bending load pressure for a fixed static time, record the real-time temperature of the medium tank, the medium ion concentration, the initial potential of the electrode, and other real-time operating parameters. Integrate all operating parameters and pipe foundation size parameters to generate an initial coupled loading test sample.

[0013] The working principle and effects of the above technical solution are as follows: By standardizing the preparation of soil corrosion media and controlling various physicochemical indicators, the detection environment can be kept uniform and stable, reducing detection errors caused by uneven media conditions. By regularly arranging a distributed reference electrode array and performing proper pretreatment of the pipe surface, the electrode acquisition positions can be made to fit the pipe structure characteristics, improving the fit and uniformity of signal acquisition throughout the process. By constructing a complete soil electrolyte three-electrode system, the stable and smooth transmission of electrochemical signals can be ensured, reducing signal interference problems during the detection process. By simultaneously applying stray current and constant bending load, the multi-physics coupling environment under actual pipe working conditions can be highly replicated, which not only fits the real service scenario but also gives the initialized test samples higher detection reference value, effectively improving the authenticity and stability of subsequent corrosion detection data.

[0014] In one embodiment of the present invention, S2 includes: S21. Move the entire medium tank storing the initial coupled loading test sample into the electromagnetic shielding test chamber. Eliminate external power frequency electromagnetic field radio frequency signal noise interference inside the shielding chamber. Connect the AC excitation output line of the electrochemical workstation to all electrode ports of the three-electrode system and lock the line contact resistance value to remain constant. S22, Set the upper and lower limits of the workstation scanning frequency respectively. Hertz and Hertz, divide the logarithmically uniform stepped frequency acquisition nodes, set the single-node signal excitation amplitude and single signal acquisition dwell time, and start the continuous scanning program to deliver broadband AC excitation signal to the pipe surface; S23. Simultaneously capture the four types of feedback values ​​of impedance magnitude, phase shift, real part impedance, and imaginary part impedance at each frequency node, continuously store the feedback values ​​corresponding to all frequency nodes, fully collect all feedback values ​​in the entire frequency band, and summarize to obtain the original full-band impedance response data without noise reduction. S24. Retrieve the preset mathematical model of the equivalent circuit of the coated metal interface, import all the original data of impedance response across the entire frequency band, carry out iterative fitting calculation, correct the parameters of the equivalent components inside the model round by round, and reduce the error difference between the calculated curve and the measured response curve until the error is lower than the allowable threshold. S25. Separate the three types of values ​​that continuously change with the corrosion process from the equivalent circuit model after fitting and convergence: coating porosity, interface peeling degree, and double layer capacitance. Bind the scanning acquisition time corresponding to each set of values, sort them in chronological order, and combine them to generate a standardized impedance characteristic parameter dataset.

[0015] The working principle and effects of the above technical solution are as follows: Impedance testing in an electromagnetically shielded environment isolates signal disturbances caused by various external electromagnetic clutter, maintains stable contact of the testing line, and reduces the adverse effects of the external environment on the testing signal. The wide-band stepped frequency scanning acquisition method fully covers the response frequency band related to pipe corrosion, compensating for data gaps in narrow-band scanning and improving the comprehensiveness of impedance data acquisition. The complete capture and retention of multi-dimensional impedance values ​​preserves original testing details, providing sufficient data support for subsequent data fitting calculations. The iterative fitting and correction of model parameters continuously reduces the deviation between simulated and measured curves, effectively improving data fitting accuracy. This approach accurately extracts core corrosion parameters of the pipe coating and interface while ensuring the regularity and accuracy of the final impedance characteristic data, providing a reliable data foundation for subsequent corrosion characteristic analysis.

[0016] In one embodiment of the present invention, step S24 includes: Retrieve the mathematical model of the equivalent circuit of the coated metal interface built into the device's background, clear the historical calculation parameters and residual data inside the model, complete the initialization of the model's calculation parameters, and generate a blank iterative calculation model. The obtained full-band impedance response raw data are imported in batches into the blank iterative calculation model to complete the adaptation and access of the measured data and the model calculation port, and generate the data source to be calculated by the model. The iterative computation model is driven to perform multiple rounds of closed-loop fitting operations, synchronously updating the operating parameters of various equivalent components such as resistors, capacitors, and inductors within the model, and generating the simulated impedance response curve corresponding to each round of iteration. The simulated response curve generated in each iteration is compared with the measured response curve collected by the equipment in real time, and the numerical deviation and trajectory deviation between the two sets of curves are continuously reduced to generate dynamic error convergence data. The model's equivalent component parameters are continuously iterated and corrected until the overall error value of the dynamic error convergence data falls into the preset acceptable range. The model fitting operation is then terminated, and the optimal equivalent circuit fitting model that has converged is generated.

[0017] The working principle and effects of the above technical solution are as follows: By initializing and cleaning the parameters of the equivalent circuit model, computational interference caused by residual data from past calculations can be eliminated, preventing the problem of fitting result deviation caused by historical data residue. By batch inputting the original measured impedance data for model calculation, it can be ensured that the original detection data fully participates in the fitting process, preventing computational distortion caused by the omission of local data. Through multiple rounds of closed-loop iterative updates to the model component parameters, it can continuously fit the actual corrosion electrical response law of the pipe, gradually reducing the deviation between the simulated curve and the measured curve. By comparing the curve error in real time and continuously correcting the parameters, the model fitting accuracy can be continuously optimized, reducing the problem of insufficient adaptability caused by fixed parameter models. This allows the circuit model to adapt to pipe detection data under different corrosion states, and ensures the fit and reliability of the fitting results, providing stable model calculation support for the accurate extraction of subsequent corrosion feature parameters.

[0018] In one embodiment of the present invention, S3 includes: S31. Using the acquisition time corresponding to each data in the impedance characteristic parameter dataset as the time reference, open the noise acquisition channel of the electrochemical workstation, cut off the external AC excitation input of the system, and retain only the potential and current signal capture function of the three-electrode system under natural corrosion state to carry out uninterrupted long-term electrochemical noise acquisition. S32. Divide the continuously acquired noise signal stream according to a uniform fixed time length, divide it into multiple independent potential fluctuation subsequences and current fluctuation subsequences, remove abnormal abrupt changes in the amplitude of the subsequences that exceed the normal fluctuation range, and retain the distortion-free complete fluctuation sequence for subsequent calculations. S33. Select a multi-layer wavelet packet basis function suitable for analyzing weak signals of metal coating corrosion, set a fixed value for the number of wavelet packet decomposition layers, input each segment of potential current fluctuation sequence into the decomposition operation module, decompose the signal layer by layer to obtain independent components corresponding to different sub-frequency bands, count the signal energy value carried by each frequency band component, and calculate the energy distribution ratio of each frequency band within a single segment sequence. S34. Select the total energy of all frequency bands in a single wave sequence as the normalization benchmark value, and convert the energy distribution ratio of each frequency band to the benchmark value to eliminate the numerical deviation caused by the difference in the total energy of the signal in different acquisition periods, and obtain the dimensionless standardized frequency band energy ratio array. S35. Based on the standardized frequency band energy ratio array, calculate the output value of the information entropy function, select the corresponding results of the entropy value matching the initial expansion interval of the microcracks on the surface of the pipe, match the wavelet packet energy entropy values ​​of each group with the impedance characteristic parameters of the corresponding time period, and integrate and package them to generate a corrosion precursor feature dataset.

[0019] The working principle and effects of the above technical solution are as follows: By shutting off external AC excitation and collecting electrochemical noise signals under natural corrosion conditions, the true microscopic corrosion reaction state of the pipe can be restored, avoiding the distortion of the original signal caused by external electrical signal interference. By equally segmenting the continuous noise signal and removing abnormal fluctuation points, invalid data caused by environmental disturbances can be filtered out, retaining pure and effective signal sequences. By processing the fluctuation signal through multi-layer wavelet packet decomposition, the signal energy characteristics of different frequency bands can be refined, significantly improving the ability to capture weak corrosion precursor signals. Through frequency band energy normalization conversion, the signal energy differences at different acquisition times can be offset, reducing the impact of data deviation on the analysis results. This solution can accurately extract corrosion feature information at the initiation stage of microcracks in the pipe, while ensuring the stability and accuracy of the corrosion precursor dataset, effectively improving the identification of early latent corrosion defects.

[0020] In one embodiment of the present invention, S32 includes: Receive the continuous electrochemical noise signal stream acquired through continuous acquisition, uniformly set the standardized signal interception time, and perform equal segmentation and trimming on the entire continuous signal stream to generate multiple sets of time-continuous initial potential fluctuation subsequences and initial current fluctuation subsequences. Retrieve the standard noise amplitude fluctuation range of pipes under natural corrosion conditions stored in the equipment, match it with the current detection environment working conditions to correct the amplitude threshold of the range, and generate a real-time noise amplitude judgment standard system; The entire initial fluctuation subsequence is traversed group by group, and the range of the signal amplitude within the sequence is compared with the range of the real-time noise amplitude judgment standard system. Abnormal signal points with amplitudes exceeding the standard range are screened out, and various abnormal fluctuation data segments are marked. Signal smoothing and completion processing is carried out on the incomplete subsequence after removing outliers. Temporal interpolation algorithm is used to fill the local signal gaps to ensure the complete duration and temporal continuity of the single subsequence signal. All fluctuation data that have undergone anomaly removal and smoothing repair are organized, and the subsequence storage format and temporal arrangement order are unified to generate a lossless complete fluctuation sequence set that can directly participate in multi-level wavelet packet decomposition operations.

[0021] The working principle and effects of the above technical solution are as follows: By equally segmenting and cutting the continuous electrochemical noise signal, various fluctuating signals can maintain a uniform timing specification, eliminating the subsequent calculation imbalance caused by inconsistent signal durations. By updating the noise amplitude judgment standard in conjunction with real-time detection conditions, it can adapt to the signal judgment requirements of different detection environments, improving the adaptability of abnormal signal screening. By systematically checking and eliminating abnormal fluctuation points, invalid noise data generated by environmental disturbances and equipment fluctuations can be filtered out, reducing the interference of abnormal signals on corrosion feature analysis. By completing and repairing incomplete signals through temporal interpolation, the integrity and temporal continuity of signal data can be maintained, avoiding deviations in analysis results due to data loss. It can thoroughly purify the original detection signal, remove invalid interference information, and provide regular and reliable signal samples for subsequent wavelet packet decomposition operations, steadily improving the accuracy of early corrosion feature identification of pipe materials.

[0022] In one embodiment of the present invention, step S4 includes: S41. Retrieve the spatial coordinate storage file entered during the deployment phase of the distributed reference electrode array. The file records the axial and radial positions of the tube body corresponding to each electrode probe and the identification of the tube section to which it belongs. Establish a temporal and spatial bidirectional mapping association between the coordinate storage file and the corrosion precursor feature dataset. S42. Write the loop traversal operation logic, sequentially retrieve the impedance characteristic parameters and wavelet packet energy entropy values ​​of each group of bound electrode identifiers in the dataset, retrieve the tube body spatial coordinates matched by the corresponding electrode probe for that group of data, and establish a complete feature coordinate corresponding unit for a single acquisition node. S43. Set three sets of judgment criteria: critical porosity of coating failure, critical threshold of interface peeling, and critical threshold of microcrack entropy. Compare the feature values ​​stored at each acquisition node with the corresponding critical thresholds and mark the electrode points whose feature values ​​exceed the critical thresholds as suspected corrosion points. S44. Match the pipe section identifiers of all suspected corrosion points, distinguish the points to three special structural locations: spigot and socket section, half sleeve covering section, pipe sleeper support section, merge consecutive suspected points with adjacent spatial coordinates and synchronously exceeding the limit of characteristic values, and delineate contiguous corrosion hot spots. S45. According to the arrangement order of the pipe from the starting end to the end along the axial direction, enter the spatial coordinate structure type feature exceeding the limit of each corrosion hot spot area in sequence, arrange all hot spot area record entries in an orderly manner, and combine them to form a continuous and complete corrosion hot spot distribution sequence of the whole pipe body.

[0023] The working principle and effects of the above technical solution are as follows: By establishing a two-way mapping relationship between detection feature data and pipe spatial coordinates, precise correspondence between corrosion feature data and pipe location can be achieved, eliminating positioning deviations caused by data and point discrepancies. By traversing and comparing various corrosion feature values ​​and critical thresholds node by node, corrosion anomalies throughout the entire pipe can be meticulously investigated, reducing the probability of missing local corrosion risks. By conducting targeted identification and division of special structural sections of the pipe, corrosion problems in various structurally weak locations can be accurately captured, conforming to the corrosion patterns of the actual easily worn areas of the pipe. By integrating adjacent abnormal points to form contiguous hotspot areas, misjudgments caused by fluctuations in single-point data can be avoided, improving the rationality of corrosion area division. This not only enables refined spatial positioning of pipe corrosion defects but also organizes global corrosion distribution information, providing real and complete spatial data support for subsequent multi-source data fusion analysis and corrosion evolution prediction.

[0024] One embodiment of the present invention, such as Figure 2 As shown, S5 includes: S51. Split the entire pipe corrosion hotspot distribution sequence into each hotspot record entry, extract the spatial coordinate structure location identifier, hotspot corrosion severity, and other spatial structure dimension information contained in the entry, split the impedance feature parameter dataset to extract the coating interface electrical evolution dimension information, and split the corrosion precursor feature dataset to extract the micro-damage signal entropy evolution dimension information. S52. Using a unified timestamp as the splicing association identifier, the spatial structure information, electrical evolution information and micro-damage entropy information corresponding to the same electrode point at the same acquisition time are merged into a multi-dimensional fusion feature vector, and a complete feature vector group without missing information is generated one by one. S53. Retrieve the Bayesian multi-source fusion model that has completed sample training and converged in the early stage. The model stores the prior probability distribution function corresponding to different degrees of erosion. Input all multi-dimensional fusion feature vectors into the model data input port in batches. S54. The model assigns an independent corrosion contribution weight coefficient to each type of input feature vector, and completes the multi-source feature fusion weighted calculation based on the Bayesian probability update rule to output the posterior probability value of pipe corrosion damage at different spatial locations and times. S55. Using the axial spatial position of the pipe as the horizontal axis and the posterior probability value of corrosion damage as the vertical axis, plot the damage distribution curves corresponding to different acquisition times in chronological order. Connect all time-series curves to fully present the overall corrosion trend of the pipe over time and generate the corrosion degradation evolution curve of the entire pipe body.

[0025] The working principle and effects of the above technical solution are as follows: By splitting detection information from three different sources and unifying the timestamps to complete feature stitching, it can fill the information gaps existing in single-type data and eliminate the problem of one-sided judgments caused by relying on single parameters to evaluate corrosion status. By batch importing and fusing feature vectors into a mature Bayesian model to perform probability calculations, it can balance the representational role of different dimensions of features on corrosion evolution and reduce the probability of misjudgment caused by local abnormal data. By assigning differentiated weights to complete the multi-source data fusion calculation, it can quantify the probability level of corrosion development at each point and improve the accuracy of pipe corrosion degree assessment. By displaying the corrosion change trend of the entire pipe in series through time-series curves, it can intuitively present the overall evolution law of corrosion over time. It can not only comprehensively evaluate by integrating spatial location, electrical signals, and micro-damage characteristics, but also generate a coherent and complete corrosion evolution curve, providing comprehensive and reliable data support for pipe life prediction.

[0026] In one embodiment of the present invention, S54 includes: The system receives a batch of complete feature vector groups and performs dimensional decomposition and parsing on the spatial structure dimension, electrical evolution dimension, and micro-damage entropy dimension data contained in each feature vector to generate single-vector multi-dimensional feature component data. Standardize and unify the units of measurement for various feature items to eliminate the calculation bias caused by the difference in the numerical magnitude of data in different dimensions, and generate standardized feature items with unified units. Iterate through all standardized feature data, match appropriate erosion contribution weight coefficients for different categories of feature data, complete the weight assignment process for all feature data, and generate a weight-adapted feature dataset. The Bayesian probability update rule is invoked to perform iterative fusion calculations on the weight-adapted feature dataset, completing the superposition correction and probability update of multi-dimensional feature information, and generating temporal-spatial coupled erosion probability calculation data. All computational data are processed and output in a regularized manner. The acquisition time of each electrode point is matched with the spatial position of the pipe, and the posterior probability values ​​of pipe corrosion damage corresponding to different spatial points and different times are output in batches.

[0027] The working principle and effects of the above technical solution are as follows: By decomposing and analyzing multi-dimensional feature vector data, the computational details of various corrosion-related parameters can be refined individually, preventing feature masking problems caused by mixed calculations of multiple types of data. By unifying the dimensional standards of various feature data, computational biases caused by different data volumes can be balanced, reducing the interference of numerical differences on the fusion results. By differentially matching the corrosion contribution weights of various features, the influence of different parameters on pipe corrosion damage can be aligned, improving the fit of data fusion. By iteratively fusing and updating probability data using Bayesian methods, the coupling results of multi-source features can be continuously corrected, improving the accuracy of corrosion damage probability calculation. By regularizing and matching spatiotemporal information to output computational results, the accurate correspondence between data and detection points and time can be ensured. This approach not only takes into account the analytical value of multi-dimensional corrosion features but also effectively reduces the one-sidedness caused by single-data-dimensional evaluation, making the judgment results of pipe corrosion damage more stable and reliable.

[0028] In one embodiment of the present invention, step S6 includes: S61. The corrosion degradation evolution curve of the entire pipe body is split to obtain multiple sets of discrete time-series damage data points. The corrosion damage probability value of the acquisition time and spatial location corresponding to each set of data points is extracted. The evolution stages are divided into three stages: early stage low-speed corrosion, middle stage accelerated corrosion, and late stage severe corrosion. The damage growth rate of each stage is statistically analyzed. S62. Fit the corrosion damage time-series growth function for each of the three evolution stages, substitute the time-series data collected from different spatial points into the corresponding stage growth function, and extrapolate to calculate the target time corresponding to the damage probability approaching the critical damage threshold of pipe failure. S63. Compare the target duration values ​​obtained from the simulation of all spatial points of the pipe. The point with the smallest value is determined as the weakest damage node in the global pipe. Simultaneously extract the axial and radial spatial coordinates matched by this node as the spatial coordinates of the weakest node in the global pipe. S64. Combine the total service life of the pipe material with the standard design, subtract the current corrosion loading time from the critical failure target time to obtain the remaining service life prediction value of the corresponding point of the cable protection pipe to be tested, and store the prediction values ​​of all points synchronously. S65. Summarize the predicted remaining service life of all spatial points in the pipe body, the spatial coordinates of the weakest section, the corrosion growth rate of each stage, the corrosion degradation curve of the whole path, and the original multi-source detection features. These are then uniformly organized into standardized data entries, packaged and stored to generate a complete dataset of corrosion resistance test results.

[0029] The working principle and effects of the above technical solution are as follows: By splitting the corrosion evolution curve and classifying the stage-based corrosion patterns, the damage growth characteristics of the pipe at different stages can be accurately captured, reducing the calculation distortion caused by uniform function extrapolation. By fitting a dedicated time-series growth function piecewise and performing data extrapolation, the time node for each spatial point to reach the failure state can be accurately predicted, improving the precision of lifespan extrapolation. By comparing the failure duration of each point across the entire domain to screen for structurally weak locations, the key areas of the pipe most prone to failure can be accurately identified, avoiding the omissions caused by uniform overall assessment. By calculating the remaining service life of each point and storing it uniformly, the service status of the entire pipe domain can be fully covered, improving the comprehensiveness of lifespan assessment. By integrating multiple types of test data to form a standardized result set, the test output specifications can be unified, which can not only fully present the entire process of pipe corrosion degradation, but also ensure the integrity and reference value of the final test results, improving the overall accuracy of corrosion resistance testing and lifespan assessment of cable protection pipes.

[0030] In one embodiment of the present invention, S62 includes: Receive the time-series damage data corresponding to the three corrosion evolution stages obtained from the division, collect all discrete damage data points of each stage, and organize them into a stage-series damage data set. Trend fitting calculations are performed on the time-series damage data sets of each stage, and the corresponding function models of corrosion evolution are matched to generate exclusive corrosion damage time-series growth functions that are adapted to each corrosion stage. Extract time-series damage data corresponding to different spatial locations across the entire pipe area, classify and categorize the data according to the corrosion evolution stage, and generate stage-adapted time-series data groups for each spatial location. The time series data sets of the phase adaptation of all spatial points are imported one by one into the corrosion damage time series growth function of the corresponding phase, and time series extension extrapolation calculations are carried out to generate continuous damage evolution trend data for each point. By comparing continuous damage evolution trend data with the critical damage threshold for pipe failure, the time node when the damage evolution process reaches the critical threshold is tracked, and the failure target duration data corresponding to each spatial point is statistically generated.

[0031] The working principle and effects of the above technical solution are as follows: By collecting segmented time-series damage data and fitting a dedicated growth function, it can adapt to the evolution patterns of different corrosion stages of the pipe, overcoming the problem that a unified function cannot match the differences in corrosion stages, and effectively improving the fit of the damage trend fitting. By performing stage-based classification processing on the data of all points, the corrosion data of each region can be matched with the corresponding evolution model, reducing the inference error caused by model mismatch. By generating continuous damage data through time-series extension inference, the interval gaps of discrete detection data can be filled, and the dynamic process of corrosion development can be fully restored. By benchmarking the failure threshold and statistically analyzing the failure duration, the corrosion failure nodes of each point can be accurately captured. It can not only refine the inference accuracy of different corrosion stages, but also realize the synchronous inference of the life of the entire pipe, effectively reducing the prediction bias caused by the traditional overall assessment method, and making the life prediction results more consistent with the actual service degradation state of the pipe.

[0032] One embodiment of the present invention provides a corrosion resistance testing and system for cable protection pipes, the system comprising: One or more processors; Memory, used to store one or more programs; Wherein, when the one or more programs are executed by the one or more processors, the one or more processors are made to implement the method described in any one of the above.

[0033] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for testing the corrosion resistance of cable protection pipes, characterized in that, The method includes: S1. Construct a complete three-electrode system for soil electrolyte; introduce a periodic stray current signal of preset amplitude into the three-electrode system, and simultaneously apply a four-point bending constant load at a preset span position in the pipe body to complete the initialization of the multi-physics field collaborative loading state and generate an initial coupled loading detection sample. S2. Apply a broadband AC excitation signal to the initial coupled loading test sample to complete the full-band electrochemical impedance spectroscopy scan and collect the raw data of the full-band impedance response; perform equivalent circuit fitting calculation on the raw data of the full-band impedance response to generate an impedance characteristic parameter dataset. S3. Continuously collect electrochemical noise from the impedance characteristic parameter dataset, simultaneously extract the potential and current fluctuation sequences within a fixed time period, perform multi-level wavelet packet decomposition on the fluctuation sequences, extract the energy distribution ratio of each frequency band, normalize the energy distribution ratio of each frequency band, calculate the wavelet packet energy entropy value corresponding to the microcrack initiation stage, and generate a corrosion precursor feature dataset. S4. Perform node traversal operation on the corrosion precursor feature dataset, combine the spatial coordinate information of the distributed reference electrode array, compare the impedance characteristic parameters of each acquisition node with the wavelet packet energy entropy value, locate the corrosion hotspot areas in three special locations of the pipe body: the socket, the half sleeve, and the pipe support point, and generate the corrosion hotspot distribution sequence of the whole pipe body. S5. Multi-dimensional feature splicing is performed on the corrosion hotspot distribution sequence of the entire pipe body. The impedance feature parameters, corrosion precursor features and hotspot distribution coordinates are input into the pre-trained Bayesian multi-source fusion model to generate the corrosion degradation evolution curve of the entire pipe body. S6. Perform time-series extrapolation calculations on the corrosion degradation evolution curve of the entire pipe body, output the predicted value of the remaining service life of the cable protection pipe under test and the spatial coordinates of the weakest section, and generate a complete dataset of corrosion resistance test results.

2. The method for testing the corrosion resistance of a cable protection pipe according to claim 1, characterized in that, S1 includes: S11. Prepare multi-component composite soil corrosion medium, control the three basic physicochemical indicators of water content, salt content, acidity and alkalinity in the medium to a fixed threshold range, and uniformly fill the prepared corrosion medium into the corrosion-resistant sealed medium tank to maintain the overall static immersion time of the medium to reach the standard static immersion period. S12. Clean all floating dust, oil stains, and oxidation attachments from the outer wall of the cable protection pipe to be tested. Mark continuous sampling points at equal intervals along the longitudinal direction of the pipe. Paste insulating pads on the surface of each point. Fix a distributed reference electrode array on the surface of the pads to ensure that all electrode probes are tightly fitted to the outer wall of the pipe without gaps. S13. Strip the protective coating from the end of the cable protection pipe to expose the continuous internal metal conductive layer. Use a low-impedance conductive clamp to hold the exposed metal area. Connect the clamp to the working electrode port of the electrochemical workstation through a shielded wire. Bury a saturated calomel counter electrode at the bottom of the medium tank. Combine the completed reference electrode array to build a three-electrode system for soil electrolyte. S14. Retrieve the built-in signal output module of the electrochemical workstation, set three types of output parameters: fixed amplitude of periodic stray current, fluctuation period, and offset coefficient, continuously supply a continuous current signal to the three-electrode system, and simultaneously start the four-point bending loading device. Set a constant bending load value according to the nominal outer diameter of the pipe, and apply the load evenly to the two preset span areas of the pipe body. S15. Maintain the synchronous operation of current signal transmission and bending load pressure for a fixed static time, record the real-time temperature of the medium tank, the medium ion concentration, the initial potential of the electrode, and other real-time operating parameters. Integrate all operating parameters and pipe foundation size parameters to generate an initial coupled loading test sample.

3. The method for testing the corrosion resistance of a cable protection pipe according to claim 1, characterized in that, The S2 includes: S21. Move the entire medium tank storing the initial coupled loading test sample into the electromagnetic shielding test chamber. Eliminate external power frequency electromagnetic field radio frequency signal noise interference inside the shielding chamber. Connect the AC excitation output line of the electrochemical workstation to all electrode ports of the three-electrode system and lock the line contact resistance value to remain constant. S22. Set the upper and lower limits of the workstation scanning frequency to correspond to 10 respectively. -2 Hertz and 10 5 Hertz, divide the logarithmically uniform stepped frequency acquisition nodes, set the single-node signal excitation amplitude and single signal acquisition dwell time, and start the continuous scanning program to deliver broadband AC excitation signal to the pipe surface; S23. Simultaneously capture the four types of feedback values ​​of impedance magnitude, phase shift, real part impedance, and imaginary part impedance at each frequency node, continuously store the feedback values ​​corresponding to all frequency nodes, fully collect all feedback values ​​in the entire frequency band, and summarize to obtain the original full-band impedance response data without noise reduction. S24. Retrieve the preset mathematical model of the equivalent circuit of the coated metal interface, import all the original data of impedance response across the entire frequency band, carry out iterative fitting calculation, correct the parameters of the equivalent components inside the model round by round, and reduce the error difference between the calculated curve and the measured response curve until the error is lower than the allowable threshold. S25. Separate the three types of values ​​that continuously change with the corrosion process from the equivalent circuit model after fitting and convergence: coating porosity, interface peeling degree, and double layer capacitance. Bind the scanning acquisition time corresponding to each set of values, sort them in chronological order, and combine them to generate a standardized impedance characteristic parameter dataset.

4. The method for testing the corrosion resistance of a cable protection pipe according to claim 3, characterized in that, S24 includes: Retrieve the mathematical model of the equivalent circuit of the coated metal interface built into the device's background, clear the historical calculation parameters and residual data inside the model, complete the initialization of the model's calculation parameters, and generate a blank iterative calculation model. The obtained full-band impedance response raw data are imported in batches into the blank iterative calculation model to complete the adaptation and access of the measured data and the model calculation port, and generate the data source to be calculated by the model. The iterative computation model is driven to perform multiple rounds of closed-loop fitting operations, synchronously updating the operating parameters of various equivalent components such as resistors, capacitors, and inductors within the model, and generating the simulated impedance response curve corresponding to each round of iteration. The simulated response curve generated in each iteration is compared with the measured response curve collected by the equipment in real time, and the numerical deviation and trajectory deviation between the two sets of curves are continuously reduced to generate dynamic error convergence data. The model's equivalent component parameters are continuously iterated and corrected until the overall error value of the dynamic error convergence data falls into the preset acceptable range. The model fitting operation is then terminated, and the optimal equivalent circuit fitting model that has converged is generated.

5. The method for testing the corrosion resistance of a cable protection pipe according to claim 1, characterized in that, The S3 includes: S31. Using the acquisition time corresponding to each data in the impedance characteristic parameter dataset as the time reference, open the noise acquisition channel of the electrochemical workstation, cut off the external AC excitation input of the system, and retain only the potential and current signal capture function of the three-electrode system under natural corrosion state to carry out uninterrupted long-term electrochemical noise acquisition. S32. Divide the continuously acquired noise signal stream according to a uniform fixed time length, divide it into multiple independent potential fluctuation subsequences and current fluctuation subsequences, remove abnormal abrupt changes in the amplitude of the subsequences that exceed the normal fluctuation range, and retain the distortion-free complete fluctuation sequence for subsequent calculations. S33. Select a multi-layer wavelet packet basis function suitable for analyzing weak signals of metal coating corrosion, set a fixed value for the number of wavelet packet decomposition layers, input each segment of potential current fluctuation sequence into the decomposition operation module, decompose the signal layer by layer to obtain independent components corresponding to different sub-frequency bands, count the signal energy value carried by each frequency band component, and calculate the energy distribution ratio of each frequency band within a single segment sequence. S34. Select the sum of energy of all frequency bands of a single wave sequence as the normalization benchmark value, and convert the energy distribution ratio of each frequency band to the benchmark value to obtain a dimensionless standardized frequency band energy ratio array. S35. Based on the standardized frequency band energy ratio array, calculate the output value of the information entropy function, select the corresponding results of the entropy value matching the initial expansion interval of the microcracks on the surface of the pipe, match the wavelet packet energy entropy values ​​of each group with the impedance characteristic parameters of the corresponding time period, and integrate and package them to generate a corrosion precursor feature dataset.

6. The method for testing the corrosion resistance of a cable protection pipe according to claim 1, characterized in that, The S4 includes: S41. Retrieve the spatial coordinate storage file entered during the deployment phase of the distributed reference electrode array. The file records the axial and radial positions of the tube body corresponding to each electrode probe and the identification of the tube section to which it belongs. Establish a temporal and spatial bidirectional mapping association between the coordinate storage file and the corrosion precursor feature dataset. S42. Write the loop traversal operation logic, sequentially retrieve the impedance characteristic parameters and wavelet packet energy entropy values ​​of each group of bound electrode identifiers in the dataset, retrieve the tube body spatial coordinates matched by the corresponding electrode probe for that group of data, and establish a complete feature coordinate corresponding unit for a single acquisition node. S43. Set three sets of judgment criteria: critical porosity of coating failure, critical threshold of interface peeling, and critical threshold of microcrack entropy. Compare the feature values ​​stored at each acquisition node with the corresponding critical thresholds and mark the electrode points whose feature values ​​exceed the critical thresholds as suspected corrosion points. S44. Match the pipe section identifiers of all suspected corrosion points, distinguish the points to three special structural locations: spigot and socket section, half sleeve covering section, pipe sleeper support section, merge consecutive suspected points with adjacent spatial coordinates and synchronously exceeding the limit of characteristic values, and delineate contiguous corrosion hot spots. S45. According to the arrangement order of the pipe from the starting end to the end along the axial direction, enter the spatial coordinate structure type feature exceeding the limit of each corrosion hot spot area in sequence, arrange all hot spot area record entries in an orderly manner, and combine them to form a continuous and complete corrosion hot spot distribution sequence of the whole pipe body.

7. The method for testing the corrosion resistance of a cable protection pipe according to claim 1, characterized in that, The S5 includes: S51. Split the entire pipe corrosion hotspot distribution sequence into each hotspot record entry, extract the spatial coordinate structure location identifier, hotspot corrosion severity, and other spatial structure dimension information contained in the entry, split the impedance feature parameter dataset to extract the coating interface electrical evolution dimension information, and split the corrosion precursor feature dataset to extract the micro-damage signal entropy evolution dimension information. S52. Using a unified timestamp as the splicing association identifier, the spatial structure information, electrical evolution information and micro-damage entropy information corresponding to the same electrode point at the same acquisition time are merged into a multi-dimensional fusion feature vector, and a complete feature vector group without missing information is generated one by one. S53. Retrieve the Bayesian multi-source fusion model that has completed sample training and converged in the early stage. The model stores the prior probability distribution function corresponding to different degrees of erosion. Input all multi-dimensional fusion feature vectors into the model data input port in batches. S54. The model assigns an independent corrosion contribution weight coefficient to each type of input feature vector, and completes the multi-source feature fusion weighted calculation based on the Bayesian probability update rule to output the posterior probability value of pipe corrosion damage at different spatial locations and times. S55. Using the axial spatial position of the pipe as the horizontal axis and the posterior probability value of corrosion damage as the vertical axis, plot the damage distribution curves corresponding to different acquisition times in chronological order. Connect all time-series curves to fully present the overall corrosion trend of the pipe over time and generate the corrosion degradation evolution curve of the entire pipe body.

8. The method for testing the corrosion resistance of a cable protection pipe according to claim 7, characterized in that, S54 includes: The system receives a batch of complete feature vector groups and performs dimensional decomposition and parsing on the spatial structure dimension, electrical evolution dimension, and micro-damage entropy dimension data contained in each feature vector to generate single-vector multi-dimensional feature component data. Standardize and unify the units of measurement for various feature items to eliminate the calculation bias caused by the difference in the numerical magnitude of data in different dimensions, and generate standardized feature items with unified units. Iterate through all standardized feature data, match appropriate erosion contribution weight coefficients for different categories of feature data, complete the weight assignment process for all feature data, and generate a weight-adapted feature dataset. The Bayesian probability update rule is invoked to perform iterative fusion calculations on the weight-adapted feature dataset, completing the superposition correction and probability update of multi-dimensional feature information, and generating temporal-spatial coupled erosion probability calculation data. All computational data are processed and output in a regularized manner. The acquisition time of each electrode point is matched with the spatial position of the pipe, and the posterior probability values ​​of pipe corrosion damage corresponding to different spatial points and different times are output in batches.

9. The method for testing the corrosion resistance of a cable protection pipe according to claim 1, characterized in that, The S6 includes: S61. The corrosion degradation evolution curve of the entire pipe body is split to obtain multiple sets of discrete time-series damage data points. The corrosion damage probability value of the acquisition time and spatial location corresponding to each set of data points is extracted. The evolution stages are divided into three stages: early stage low-speed corrosion, middle stage accelerated corrosion, and late stage severe corrosion. The damage growth rate of each stage is statistically analyzed. S62. Fit the corrosion damage time-series growth function for each of the three evolution stages, substitute the time-series data collected from different spatial points into the corresponding stage growth function, and extrapolate to calculate the target time corresponding to the damage probability approaching the critical damage threshold of pipe failure. S63. Compare the target duration values ​​obtained from the simulation of all spatial points of the pipe. The point with the smallest value is determined as the weakest damage node in the global pipe. Simultaneously extract the axial and radial spatial coordinates matched by this node as the spatial coordinates of the weakest node in the global pipe. S64. Combine the total service life of the pipe material with the standard design, subtract the current corrosion loading time from the critical failure target time to obtain the remaining service life prediction value of the corresponding point of the cable protection pipe to be tested, and store the prediction values ​​of all points synchronously. S65. Summarize the predicted remaining service life of all spatial points in the pipe body, the spatial coordinates of the weakest section, the corrosion growth rate of each stage, the corrosion degradation curve of the whole path, and the original multi-source detection features. These are then uniformly organized into standardized data entries, packaged and stored to generate a complete dataset of corrosion resistance test results.

10. A corrosion resistance testing and system for cable protection pipes, characterized in that, The system includes: One or more processors; Memory, used to store one or more programs; Wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1 to 9.

Images