A heat supply pipeline corrosion state monitoring and protection early warning system

Abstract

The present application belongs to the technical field of pipeline corrosion detection, and discloses a heat supply pipeline corrosion state monitoring and protection early warning system. Through the same-section alignment sensing layout architecture, the corrosion correlation data of the pipeline body, the conveying medium and the buried environment can be synchronously acquired. Through the feature-level precise registration algorithm based on the corrosion physical mechanism, the detection features are uniformly mapped to the corrosion rate physical dimension. Through the evaluation framework of corrosion mechanism analysis and data-driven fusion, the identification of four typical corrosion types is automatically completed. The rate, damage parameter and residual pressure-bearing capacity of corrosion are calculated. Combined with the time series prediction model embedded with physical constraints, the residual wall thickness distribution and residual service life probability interval of the pipeline are output. The corrosion state of the pipeline can be quantified, and the abnormal pipe section with accelerated corrosion can be identified. According to the real-time corrosion evaluation result, the operation parameters of the cathodic protection, the inhibitor addition, the water discharge and the stray current drainage device are adjusted.

Description

Technical Field

[0001] This invention belongs to the field of pipeline corrosion detection technology, specifically a system for monitoring and protecting the corrosion status of heating pipelines. Background Technology

[0002] Heating pipelines transport high-temperature and high-pressure heat media over long periods, and most are buried underground. The pipe walls simultaneously face multiple corrosive effects, including oxygen corrosion and under-deposit corrosion from the internal media, as well as electrochemical corrosion and stray current corrosion from the external soil. Corrosion leads to uniform thinning of the pipe wall and localized pitting and perforation, which is a core cause of leaks and pipe bursts in heating pipelines. Current technologies for monitoring and protecting against corrosion in heating pipelines still face the following technical challenges:

[0003] Traditional manual inspections and periodic ultrasonic testing are not timely and cannot achieve continuous monitoring of the entire pipeline section around the clock; multi-source corrosion detection data lacks a unified spatiotemporal benchmark and physical dimension alignment, and multi-dimensional detection features cannot be directly integrated for accurate characterization of corrosion status; corrosion assessment relies on simple threshold judgments and does not incorporate multi-field coupling factors and corrosion physical laws constraints; corrosion type identification relies on manual experience judgment; the automatic differentiation of four typical corrosion types is not accurate enough and cannot support differentiated corrosion control.

[0004] Early warning systems generally use fixed threshold standards and do not dynamically adjust them based on the importance of the pipe section and the corrosion evolution trend, resulting in weak differentiated graded early warning capabilities.

[0005] The monitoring and protection processes are disconnected, and the system is not linked with cathodic protection, corrosion inhibitor dosing and other devices in a closed loop. The protection strategy is crude and prone to problems of insufficient or excessive protection. At the same time, corrosion monitoring data is stored in a scattered manner, lacking a full-cycle quantitative assessment mechanism for corrosion evolution, an offline benchmark calibration mechanism and a full-process digital management and control system, making it difficult to support pipeline corrosion evolution analysis and full life cycle operation and maintenance decisions. Summary of the Invention

[0006] The purpose of this invention is to provide a system for monitoring and protecting the corrosion status of heating pipelines, so as to solve one or more problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a corrosion status monitoring and protection early warning system for heating pipelines, comprising a data monitoring module, a processing and storage module, a status assessment module, a graded early warning module, a protection effectiveness verification module, a testing and calibration module, and a detection and judgment module;

[0008] Furthermore, in the data monitoring module, distributed optical fiber sensing units are arranged along the outer wall of the heating pipeline to synchronously collect axial strain and temperature distribution data of the pipe wall, and identify stress concentration areas caused by pipe wall corrosion thinning through abnormal strain distribution.

[0009] At key nodes, the detection units are arranged in the same cross section alignment manner. The ultrasonic phased array thickness measurement probe and the electrochemical impedance spectroscopy sensor are arranged at the corresponding positions on the outer wall of the same circumferential cross section of the pipe. The ultrasonic phased array thickness measurement probe uses the pulse reflection method to detect the remaining wall thickness of the pipe wall, and at the same time collects the defect echo signal to identify the morphological characteristics of pitting corrosion and ulcer corrosion.

[0010] During identification, three types of feature parameters are extracted from the defect echo: amplitude distribution, propagation time difference, and echo waveform distortion. The extracted feature parameters are compared one by one with a preset standard corrosion morphology template library to match the morphology type with the highest echo feature similarity. The axial length and circumferential width of the corrosion defect are determined according to the spatial distribution range of the echo, and the specific type and geometric boundary parameters of the corrosion morphology are marked.

[0011] The electrochemical impedance spectroscopy sensor is embedded in the outer wall of the pipe at the contact surface with the soil. By applying a sinusoidal potential perturbation, it collects the impedance spectrum characteristics of the electrode system and calculates the corrosion current density. Insertion-type dissolved oxygen sensors and conductivity sensors are arranged at corresponding cross sections inside the pipe to collect the dissolved oxygen content and ionic conductivity of the hot water inside the pipe. Soil parameter sensors are arranged at intervals along the pipe to collect the pH value, moisture content, chloride ion content, redox potential, and stray current density of the buried soil.

[0012] Furthermore, in the processing and storage module, a signal feature-level fine registration algorithm based on corrosion physics priors is adopted. First, based on the temperature-corrosion rate Arrhenius relationship, the monotonically decreasing wall thickness law, and the stress-wall thickness elasticity formula, data rationality verification rules are constructed to eliminate outliers caused by sensor failures. Then, through a linear interpolation time alignment algorithm, sensor data with different sampling frequencies are unified to the same time reference. Combined with the pipeline GIS coordinate system and the layout rules of alignment with the same cross section, the detection data is mapped to the axial mileage and circumferential cross section coordinates of the pipeline, completing the coarse alignment in both time and space dimensions.

[0013] After coarse alignment, corresponding denoising algorithms are used to process the noise characteristics of different detection data. Finally, based on Faraday's law, stress-wall thickness elasticity formula and other corrosion physical mechanisms, and based on Faraday's law, the correspondence between wall thickness reduction rate and stress concentration accelerating corrosion mechanism, the three types of features, namely corrosion current density of electrochemical impedance spectroscopy, wall thickness reduction amount of ultrasonic thickness measurement, and stress change amount calculated by optical fiber strain, are uniformly mapped to the dimension of corrosion rate physical quantity to complete feature-level registration.

[0014] The detection parameters of different dimensions are normalized, the raw and preprocessed data of high-frequency monitoring are stored in a time-series database, and the corrosion assessment results and early warning records of key nodes are stored using blockchain technology.

[0015] Furthermore, in the state assessment module, a corrosion mechanism feature fingerprint database is constructed. Through multi-feature matching, the identification and judgment of four typical corrosion types are automatically completed. For uniform electrochemical corrosion, based on Faraday's law and combined with the corrosion current density obtained by electrochemical impedance spectroscopy, the calculation process simultaneously superimposes the Arrhenius correction term for the temperature of the medium inside the pipe and the accelerated corrosion correction term for the pipe wall stress to obtain the uniform corrosion rate and cumulative corrosion thinning of the pipeline; combined with the original wall thickness of the pipeline, the remaining wall thickness is obtained.

[0016] For local pitting corrosion, based on ultrasonic phased array defect echo imaging, the maximum depth of pitting corrosion, pit area, and local thinning rate are calculated using a pre-calibrated stress-pitting extended finite element simulation inversion model. At the same time, the remaining pressure bearing capacity of the corroded pipe section is calculated by combining the yield strength of the pipe material.

[0017] For under-deposit corrosion, the under-deposit corrosion rate and the degree of interface damage are calculated by combining the scale thickness and interface corrosion characteristics detected by ultrasonic testing and superimposing the under-deposit occlusion cell corrosion acceleration coefficient; for stray current corrosion, the AC corrosion rate and pitting corrosion propagation probability are calculated by combining the coupling relationship between stray current density and corrosion current density.

[0018] Before calculation, retrieve the stray current density time series data of the corresponding pipe section and the corrosion current density benchmark value obtained by electrochemical detection, and extract the characteristic parameters of stray current amplitude, frequency and duration; match the preset stray current-corrosion coupling coefficient reference table, and select the coupling coefficient value under the corresponding working condition.

[0019] The AC corrosion rate is calculated by taking the uniform corrosion rate corresponding to the natural corrosion current density as the base value and combining it with the coupling coefficient; the frequency of stray current peak occurrence within a set period is statistically analyzed, and the pitting corrosion propagation probability within the corresponding period is calculated by combining it with the critical condition for pitting corrosion initiation.

[0020] A time-series prediction model integrating corrosion mechanism constraints is constructed. The model takes historical corrosion data, operating parameters, and environmental parameters of the pipeline as inputs, and uses Faraday's law, the Arrhenius temperature effect formula, and the stress-accelerated corrosion formula as physical constraints. The model is embedded in a long short-term memory network model in the form of a loss function regularization term to predict the distribution range of the remaining wall thickness of the pipeline at different future time points. Combined with the pipeline safe wall thickness threshold, the probability range of the remaining service life is calculated. At the same time, the model outputs the changing trend of corrosion development rate and identifies abnormal pipe sections with accelerated corrosion.

[0021] During calculation, the remaining wall thickness distribution range output by the time series prediction model is compared with the pipeline safe wall thickness threshold at each time step, and the corresponding time node when the remaining wall thickness first falls below the safe threshold is marked; multiple sets of different initial parameters and environmental fluctuation conditions are input, and the wall thickness prediction process is repeated to obtain multiple sets of independent lifetime prediction results.

[0022] The frequency of wall thickness exceeding the standard at each time point in the statistical prediction results is used to fit the probability distribution curve of the remaining service life; the 95% confidence interval of the probability distribution is taken as the final output probability interval of the remaining service life, and the median value of the probability distribution is output simultaneously as the reference service life value.

[0023] Furthermore, in the graded early warning module, a dynamic risk weight calculation system for the probability of accelerated corrosion evolution is established. This system integrates static attributes such as pipeline grade, operating pressure, laying area, and surrounding sensitive facilities, combined with the current corrosion type, environmental parameter change trends, and corrosion rate increments. The probability of accelerated corrosion occurring in the pipe section in the future is calculated through a corrosion mechanism model, and the risk weight coefficient of the pipe section is updated. Based on the general safe wall thickness threshold and corrosion rate threshold, the early warning trigger threshold of the pipe section is adjusted in combination with the dynamic risk weight coefficient. Higher-risk pipe sections are subject to stricter early warning standards.

[0024] During the calculation, continuous corrosion rate data of the target pipe section over the past six months were extracted, and the trend line of corrosion rate change and historical fluctuation range were fitted. Combining the seasonal variation pattern of environmental parameters and future change predictions, the corrosion mechanism model was input to obtain the corrosion rate amplification coefficient under different scenarios. The corrosion rate evolution process under multiple sets of different environmental conditions in the next 1-3 years was simulated, and the proportion of scenarios with corrosion rates exceeding twice the benchmark value was statistically analyzed to obtain the probability of accelerated corrosion of the pipe section in the corresponding period.

[0025] The system sets up four warning levels: blue, yellow, orange, and red. It also sets up three judgment dimensions: corrosion rate level and increment, absolute corrosion thinning amount, and remaining service life. Meeting the threshold corresponding to any dimension will trigger a four-level warning. Each warning level is configured with a quantitative threshold for the corresponding dimension. The blue warning is when the corrosion rate is slightly higher than the industry benchmark value under the same working conditions and the rate increment is on the rise. The yellow warning is when the local pitting depth exceeds 10% of the wall thickness. The orange warning is when the remaining wall thickness is less than 80% of the design wall thickness. The red warning is when the remaining wall thickness is close to the safe wall thickness threshold or there is deep pitting corrosion with the risk of perforation.

[0026] When an early warning is triggered, the location of the corroded pipe section is automatically located, and the corrosion type, cause analysis, and risk impact range are output simultaneously and pushed to the corresponding maintenance personnel. At the same time, the early warning level can be adjusted, and the early warning level can be changed in real time according to the monitoring data. The cause analysis includes excessive chloride ions in the soil and excessive dissolved oxygen in the pipe.

[0027] Furthermore, in the protection effectiveness verification module, for the impressed current cathodic protection device for buried pipelines, when the evaluation finds that the corrosion rate of the outer wall of the pipe section is increasing and the protection potential deviates from the optimal range, the output current of the corresponding potentiostat is adjusted to stabilize the pipe-to-ground potential relative to the saturated copper sulfate reference electrode in the optimal protection range of -0.85V to -1.10V.

[0028] After adjusting the protection parameters, the electrochemical impedance spectroscopy and power-off potential data of the corresponding pipe section are collected to eliminate the influence of IR drop. The corrosion rate inhibition rate of cathodic protection is quantitatively calculated based on the proportion of the corrosion rate difference before and after the adjustment of the protection parameters. The actual effect of cathodic protection is evaluated, the control parameters are corrected, and the risk of hydrogen embrittlement caused by overprotection is avoided.

[0029] The IR drop elimination method is adopted by instantaneous potential acquisition during power failure. First, the output current of the potentiostat is instantaneously cut off by a remote control command. Simultaneously, the potential acquisition unit is triggered to perform multiple potential acquisitions within a fixed millisecond period after power failure. After three consecutive power failure acquisitions, abnormal acquisition values ​​with deviations exceeding the preset range are eliminated. The average value of the remaining valid acquisition values ​​is taken as the final result. The ohmic voltage drop component generated between the soil medium and the reference electrode is removed to obtain the true polarization potential data of the pipeline.

[0030] The automatic corrosion inhibitor dosing device in the linkage pipe adjusts the concentration and frequency of corrosion inhibitor dosing based on the dissolved oxygen content and corrosion rate of the inner wall. After the dosing parameters are adjusted, the corrosion inhibition efficiency of the corrosion inhibitor is calculated based on the detection data of the corrosion rate of the inner wall. The control logic based on corrosion rate feedback is adopted, and the dosage is gradually reduced when the corrosion rate of the inner wall drops below the target value.

[0031] When reducing the dosage, the corrosion inhibitor concentration is gradually reduced according to a preset fixed ratio, with each reduction not exceeding 20% ​​of the current dosage. Each adjustment of the dosage parameters maintains a preset operating cycle. After the cycle ends, multiple sets of internal wall corrosion rate detection data are continuously collected, and the average value is compared with the target value. If the corrosion rate is still lower than the target value, the next round of reduction is carried out. The dosage is stopped after it drops to the preset minimum maintenance dose. If the corrosion rate rises above the target value during the reduction process, the dosage parameters are adjusted back to the previous level.

[0032] For pipe sections that are low-lying and prone to water accumulation, when the soil moisture content is detected to be continuously exceeding the standard, the surrounding water drainage device is activated. For pipe sections with excessive stray current, the stray current drainage device is activated. After the protective measures are activated, the corrosion rate of the corresponding pipe section is monitored. The protective effect is quantitatively evaluated through the detection data within a preset period. If the expected target is not achieved, the protective parameters are adjusted until the corrosion rate is controlled within the preset range.

[0033] When determining soil moisture content, the soil moisture content detection values ​​of a continuous preset number of sampling periods are statistically analyzed. If all values ​​exceed the preset activation threshold, it is determined to be a continuous exceedance, excluding false triggering caused by single data fluctuations. When determining stray current, the effective average value of stray current density within a unit time window is statistically analyzed. If the average value exceeds the preset activation threshold, the drainage device is triggered to start. After the determination is completed, an activation command is automatically sent to the corresponding protection device, and the command issuance time and the device's feedback operating status are recorded.

[0034] First, the average corrosion rate within a preset period after the protective measures are activated is statistically analyzed and compared with the baseline corrosion rate value of the same period before the activation of the protective measures. The corrosion rate reduction ratio is calculated as the protective effectiveness index. The calculated protective effectiveness index is compared with the preset compliance threshold. If the index reaches or exceeds the threshold, the protective effectiveness is deemed to be up to standard, and the current operating parameters are maintained.

[0035] If the standard is not met, the operating parameters of the corresponding protective device are adjusted upwards according to the preset fixed step size. After the parameter adjustment is completed, wait for the preset stabilization period. After the protective effect is fully effective, collect test data again to evaluate the protective effectiveness. Repeat the parameter adjustment and effectiveness verification process until the protective effectiveness meets the standard or the maximum number of adjustments is reached.

[0036] Furthermore, in the test and calibration module, the pipeline corrosion status visualization and control interface displays the corrosion degree, remaining service life, and warning level of the pipe section in the form of a heat map. It supports pipe section drilling. Clicking on the pipe section allows you to view real-time sensor data, historical corrosion trends, evaluation reports, warning records, and protection effectiveness verification. It also displays the operating status of monitoring equipment and protection equipment.

[0037] A service life corrosion evolution map and benchmark calibration system were constructed, integrating material corrosion performance test data during pipeline construction, regular offline excavation test data during service, and online continuous monitoring data. Each offline benchmark test result and excavation verification data after emergency response were used as calibration values ​​to iteratively calibrate the parameters and early warning thresholds of the online evaluation model.

[0038] When constructing the map, a two-dimensional coordinate system is established with the pipeline axial mileage as the horizontal axis and the service time as the vertical axis, and the spatial resolution and temporal resolution are set. Data such as corrosion rate, remaining wall thickness, early warning records, and protective measures records of each historical stage are mapped to the corresponding coordinate points of the coordinate system. Spatial interpolation algorithms are used to fill in the parameter values ​​of unmonitored pipe sections to generate a continuous spatiotemporal distribution map of corrosion parameters. The map is divided into multiple independent layers according to parameter type, and supports layer overlay viewing and single layer switching. The map content of the corresponding area is automatically updated after new monitoring data is added.

[0039] During calibration, the time segment and pipe section location corresponding to the offline calibration data are extracted first, and the corresponding output data of the online evaluation model at the same time are retrieved. The offline calibration values ​​and the online model output values ​​are compared group by group, and the deviation values ​​and deviation rates of the three core parameters of corrosion rate, remaining wall thickness and remaining service life are calculated respectively. The correction coefficients and early warning trigger thresholds of the corresponding calculation modules in the model are adjusted proportionally according to the deviation values.

[0040] After the correction is completed, import the historical monitoring dataset of the past year for backtesting and verification to confirm that the deviation rate of the output data after calibration is within the preset allowable range. After the verification is passed, update the model's running parameter library and early warning threshold configuration table, and synchronously record the calibration data source and parameter change records of this calibration.

[0041] It has a built-in hidden danger work order management system. After the warning is triggered, it automatically generates a hidden danger handling work order and dispatches it to the corresponding operation and maintenance team. It tracks the handling progress and results of the work order, supports the formulation of inspection plans and the uploading of inspection records, and integrates manual inspection data with online monitoring data.

[0042] Before merging, the paper or electronic data of manual inspection records are first converted into a structured form, extracting core fields such as corrosion detection values, equipment operating status, and on-site anomaly records, and matching them with the unique number and inspection timestamp of the corresponding pipe section. The converted inspection data is then aligned with the online monitoring data of the same time period and pipe section in terms of time dimension, and the axial mileage of the pipeline is used as the spatial coordinate benchmark to calibrate the spatial positional deviation between the two types of data.

[0043] The parameters of the two types of data are uniformly converted, and the same physical quantities are given uniform units and data precision. The data are then organized into data entries with a uniform format according to a standardized field structure. The aligned data are jointly stored in the historical corrosion database of the corresponding pipe section, and simultaneously incorporated into the corrosion trend analysis dataset and model calibration sample set. The data source is marked as manual inspection or online monitoring.

[0044] The system statistically analyzes the corrosion rate distribution, number of potential hazards, frequency of early warnings, and effectiveness of protective measures for pipelines of different regions, materials, and service life. It aims to uncover the patterns and main causes of corrosion, verify the universality of corrosion mechanism models, and generate periodic reports. The reports include a summary of the current corrosion status, trend analysis, suggestions for hazard mitigation, and operation and maintenance work plans.

[0045] Furthermore, in the detection and assessment module, when a red alert is triggered and the system assessment determines that there is a risk of leakage or pipe burst, the encrypted detection and assessment mode is activated. The distributed optical fiber sensing unit is upgraded to a high-frequency continuous sampling mode, the ultrasonic phased array probe is switched to a continuous sampling mode, and the sampling frequency of the electrochemical sensor is encrypted to 1 time / minute. The defect propagation rate and the probability of perforation risk are calculated through the defect propagation inversion model, and the early warning information, including the location of the hidden danger, the risk level, the expected impact range, and preliminary handling suggestions, is pushed to the emergency management department and the operation and maintenance repair team simultaneously.

[0046] After the mode start command is generated, sampling frequency adjustment commands are sent sequentially to the distributed fiber optic sensing unit, ultrasonic phased array thickness probe, and electrochemical impedance spectroscopy sensor according to device priority, setting the high-frequency sampling parameters for the corresponding devices. After the commands are sent, the operating status parameters returned by each device are verified. Once it is confirmed that the device has completed the parameter switching and is operating normally, high-frequency data acquisition and real-time analysis are officially started. During the encrypted mode operation, a device status check is performed at fixed intervals to verify the sampling frequency and data transmission status. An alarm is automatically triggered when a device goes offline or data anomalies occur.

[0047] After the emergency response is completed and the red alert is lifted, a mode recovery command is generated, and a sampling frequency callback command is issued to each device to restore the sampling parameters of each device to the normal operating configuration. After confirming that the device has returned to normal operating status, the encrypted detection and analysis mode is officially exited, and the duration of encrypted mode operation and the amount of data collected are counted simultaneously.

[0048] When calculating the expected impact range, the pipeline topology data and valve layout ledger are retrieved. Combined with the location, diameter, and upstream and downstream valve locations of the potentially hazardous pipe section, a topology traversal algorithm is used to simulate the medium flow path after the potentially hazardous pipe section is cut off. Downstream isolated branch pipe sections and covered users are investigated segment by segment. The total length of affected pipe sections, user types and numbers are counted. At the same time, the geographical coordinate range of the affected area is marked to generate standardized risk impact range data.

[0049] First, the continuously high-frequency collected wall thickness detection data, corrosion rate data, and pipeline operating pressure data are time-series aligned to extract the initial depth, pit area, location information, and other dimensional parameters of the current defect. Then, combined with the fracture toughness parameters and yield strength parameters of the pipeline material and the acceleration coefficient of the corrosion environment, the expansion and change process of the defect size is gradually deduced step by step at fixed small time steps. During the deduction process, the stress alternation effect caused by the operating pressure fluctuation is simultaneously superimposed, and the depth expansion and area expansion of the defect are calculated step by step. The real-time expansion rate of the defect and the probability of perforation within the set prediction time are output.

[0050] The system is linked to the heating network dispatching system to adjust the operating parameters of the corresponding pipe section. When necessary, the system remotely controls the shut-off valve to cut off the medium transport in the potentially hazardous pipe section. At the same time, the system activates the backup pipeline to ensure downstream heating. The system is linked to the geographic information system to retrieve information on underground pipelines, valves, manholes and other facilities around the potentially hazardous pipe section. The system is linked to the emergency command system to simultaneously push monitoring data from the accident site.

[0051] After the emergency response is completed, the system automatically records the detection data of the emergency process, including the development process of the hidden danger, the warning time, the response process, and the response results. Combined with the subsequent excavation and detection results, the corrosion assessment model and the warning threshold are corrected.

[0052] The beneficial effects of this invention are as follows:

[0053] 1. This invention, through a sensing deployment architecture aligned with the same cross section, can simultaneously acquire corrosion correlation data of the pipeline body, the transport medium, and the buried environment. Using a feature-level fine registration algorithm based on the physical mechanism of corrosion, the detection features are uniformly mapped to the physical dimension of corrosion rate. Through an evaluation framework that integrates corrosion mechanism analysis and data-driven fusion, it automatically completes the identification of four typical corrosion types, calculates the corrosion rate, damage parameters, and remaining pressure bearing capacity, and outputs the probability range of the pipeline's remaining wall thickness distribution and remaining service life by combining a time-series prediction model with embedded physical constraints. This enables the quantification of pipeline corrosion status and the identification of abnormal pipe sections with accelerated corrosion.

[0054] 2. Based on real-time corrosion assessment results, this invention adjusts the operating parameters of cathodic protection, corrosion inhibitor addition, water discharge, and stray current drainage devices accordingly. By quantitatively calculating the actual effectiveness of protective measures through continuous corrosion detection data, and iteratively optimizing the protective parameters with corrosion rate changes as feedback indicators, the protective intensity can be stably controlled within a reasonable range, ensuring the protective effect while avoiding additional damage caused by overprotection, so that the pipeline corrosion rate is maintained within the preset safe range for a long time.

[0055] 3. This invention establishes a four-level graded early warning mechanism that combines dynamic risk weights of pipe sections. It triggers corresponding early warning levels through multi-dimensional criteria and outputs differentiated early warning standards to adapt to the risk attributes of different pipe sections. It realizes intuitive presentation of the corrosion status of the pipeline network and full-dimensional data drilling through a visual management and control interface, integrates full-cycle corrosion data to build a benchmark calibration system, and continuously iterates and optimizes the evaluation model and early warning thresholds. In high-risk scenarios, it automatically starts the encrypted detection and judgment mode, links with external business systems to support emergency response, and improves the full-cycle management and control capabilities of corrosion hazards in heating pipeline networks. Attached Figure Description

[0056] Figure 1 This is a flowchart of the corrosion monitoring and early warning process for heating pipelines according to the present invention;

[0057] Figure 2 This is a flowchart of the corrosion status assessment process of the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] like Figures 1 to 2As shown, this embodiment of the invention provides a system for monitoring and protecting the corrosion status of heating pipelines, including a data monitoring module, a processing and storage module, a status assessment module, a graded early warning module, a protection effectiveness verification module, a testing and calibration module, and a detection and judgment module.

[0060] In this embodiment of the invention, the data monitoring module is equipped with distributed optical fiber sensing units arranged axially along the outer wall of the heating pipe to synchronously collect axial strain and temperature distribution data of the pipe wall. By identifying the stress concentration area caused by pipe wall corrosion and thinning through abnormal strain distribution, the spatial resolution can reach 0.5m.

[0061] Before identification, with the pipeline shut down and the medium temperature balanced with the ambient temperature, axial strain data of the entire pipe section were collected three times consecutively. After removing outliers, the average value of each point was taken to establish the strain benchmark distribution curve for the corresponding pipe section. The corresponding ambient temperature was recorded simultaneously as the benchmark temperature. During identification, the axial strain data collected in real time was first corrected for temperature to eliminate the influence of temperature changes on the strain value. The corrected strain distribution data was then matched and compared with the benchmark curve point by point according to the axial mileage, and the strain deviation value of each sampling point was calculated.

[0062] A fixed-length sliding window is used to traverse sampling points along the pipeline axis. Sections where the strain deviation at three or more consecutive points within the window exceeds a preset threshold, and the deviation values ​​exhibit a localized increasing gradient distribution, are initially marked as candidate stress concentration areas. These candidate areas undergo secondary verification to exclude inherent strain anomalies caused by pipeline fittings and supports. Confirmed areas are marked as corrosion-related stress concentration areas, and the axial start and end points, maximum strain deviation value, corresponding circumferential location, and spatial distribution range of each area are recorded.

[0063] At key nodes such as pipe bends, tees, welds, low-lying areas, and pipe sections with a service life of over 10 years, the detection units are arranged in the same cross-section alignment manner. The ultrasonic phased array thickness measurement probe and the electrochemical impedance spectroscopy sensor are arranged at corresponding positions on the outer wall of the same circumferential cross-section of the pipe. The ultrasonic phased array thickness measurement probe uses the pulse reflection method to detect the remaining wall thickness of the pipe wall, with a detection accuracy of up to 0.01 mm. At the same time, it collects defect echo signals to identify the morphological characteristics of pitting corrosion and ulcer corrosion.

[0064] Before installation, clean the anti-corrosion layer and rust on the corresponding position of the outer wall of the pipe, and grind out a smooth metal contact surface; use a laser line marker to draw a closed reference positioning line along the circumference of the pipe, and align the center of the ultrasonic phased array thickness probe and the electrochemical impedance spectroscopy sensor with the reference positioning line, and control the axial installation deviation of the two types of sensors to not exceed 5cm and the circumferential angle deviation to not exceed 5°.

[0065] After installation, a standard wall thickness test block with the same material as the pipe being tested was used to perform zero-point calibration and sound velocity calibration on the ultrasonic phased array thickness probe. A standard buffer solution with a known impedance value was used to perform baseline calibration on the electrochemical impedance spectroscopy sensor. After calibration, data from the reference test block were collected three times consecutively to confirm that the detection deviation was within the preset allowable range.

[0066] The electrochemical impedance spectroscopy sensor is embedded in the outer wall of the pipe at the contact surface with the soil. By applying a sinusoidal potential perturbation, it collects the impedance spectrum characteristics of the electrode system and calculates the corrosion current density. Insertion-type dissolved oxygen sensors and conductivity sensors are arranged at corresponding cross sections inside the pipe to collect the dissolved oxygen content and ionic conductivity of the hot water inside the pipe. Soil parameter sensors are arranged at intervals along the pipe to collect the pH value, moisture content, chloride ion content, redox potential, and stray current density of the buried soil.

[0067] Sensing devices are uniformly accessed through an edge computing gateway, and the sampling frequency is configured differently as needed: once per minute for distributed optical fiber, once per hour for ultrasonic phased array thickness measurement, once per 10 minutes for electrochemical sensors, and once per 30 minutes for environmental parameters.

[0068] The edge computing gateway connects to different types of sensing devices via wired industrial Ethernet and wireless LoRa communication protocols. It performs protocol parsing and format standardization conversion on the raw data uploaded by the sensors, uniformly converting it into structured JSON data format, and simultaneously performs preliminary outlier screening and data format verification. Data that passes verification is uploaded to the cloud server in batches according to a preset upload cycle. When the network connection is interrupted, the gateway automatically caches the monitoring data in chronological order to the local solid-state storage unit. When the cached data reaches 80% of the storage capacity threshold, it automatically overwrites the oldest historical data. After the network is restored, the gateway automatically resumes the interrupted transmission, re-transmitting the cached data in chronological order from oldest to newest. During the re-transmission process, the data integrity is verified simultaneously, and missing data is automatically retransmitted.

[0069] In this embodiment of the invention, the processing and storage module employs a signal feature-level fine registration algorithm based on corrosion physics priors. First, it constructs data rationality verification rules based on the temperature-corrosion rate Arrhenius relationship, the monotonically decreasing wall thickness law, and the stress-wall thickness elasticity formula to eliminate outliers caused by sensor malfunctions. Then, through a linear interpolation time alignment algorithm, it unifies sensor data from different sampling frequencies to the same time reference. Combining the pipeline GIS coordinate system and the layout rules for alignment with the same cross section, it maps the detection data to the axial mileage and circumferential cross section coordinates of the pipeline, completing coarse alignment in both time and space dimensions.

[0070] Before mapping, a conversion table between pipeline GIS geographic coordinates and pipeline axial mileage is pre-established, and the sensor's installation geographic coordinates, circumferential installation angle, and equipment number are entered. Based on the conversion table of each sensor's installation geographic coordinates, the corresponding pipeline axial mileage value is calculated, and the sensor's detection data is marked to the corresponding mileage position. Combining the layout rules of alignment within the same cross section, the data of multiple types of sensors within the same circumferential cross section are uniformly associated with the same cross section number, completing the spatial dimension coordinate alignment of the detection data.

[0071] During verification, the continuous sampling values ​​of a single type of sensor are traversed step by step over time. Anomalies are identified based on the value range of the corresponding physical quantity and the threshold of the single-step change rate. Values ​​that exceed the value range of the physical law or whose single-step change exceeds the threshold are marked as single-point anomalies. At the same time, the associated values ​​of multiple types of sensors within the same circumferential section are compared, and the data consistency is determined based on the physical correspondence of corrosion. Values ​​that do not conform to the correspondence are marked as associated anomalies. All marked anomalies are replaced with the linear interpolation results of adjacent valid values ​​in the same sequence. When the number of consecutive anomalies exceeds 3, the data for that time period is marked as invalid and is supplemented with monitoring data of adjacent pipe sections of the same type during the same period.

[0072] After coarse alignment, corresponding denoising algorithms are used to process the noise characteristics of different detection data. Among them, wavelet threshold denoising is used for distributed optical fiber strain data, moving average denoising is used for electrochemical detection data, and pulse echo filtering denoising is used for ultrasonic thickness measurement data.

[0073] Finally, based on Faraday's law, the stress-wall thickness elasticity formula and other corrosion physical mechanisms, and based on Faraday's law, the correspondence between wall thickness reduction rate and stress concentration accelerating corrosion mechanism, the three types of characteristics—corrosion current density of electrochemical impedance spectroscopy, wall thickness reduction amount of ultrasonic thickness measurement, and stress change amount calculated by optical fiber strain—are uniformly mapped to the dimension of corrosion rate physical quantity, thus completing the feature-level registration.

[0074] When denoising distributed fiber optic strain data, the corresponding wavelet basis function is selected for multi-level wavelet decomposition. The high-frequency coefficients obtained by decomposition are threshold-quantized, and then the denoised strain data is obtained through wavelet reconstruction. When denoising electrochemical detection data, a sliding window of fixed length is set, and the average value of the sampled values ​​within the window is taken as the output value at the center of the window. The data sequence is traversed along the time axis. When denoising ultrasonic thickness measurement data, the effective passband range of the echo signal is set, and out-of-band high-frequency noise and low-frequency interference noise are filtered out, while retaining the effective echo signals of the pipe wall and defects.

[0075] During the mapping process, the corrosion current density obtained from electrochemical impedance spectroscopy is first converted into an initial value of electrochemical corrosion rate according to Faraday's law; based on the difference between two adjacent wall thickness measurements and the corresponding time interval, the initial value of corrosion rate corresponding to ultrasonic thickness measurement is calculated.

[0076] Based on the change in pipe wall stress and the stress-accelerated corrosion coefficient, the initial value of the corrosion rate corresponding to the fiber strain is calculated. Then, using the corrosion rate value obtained by ultrasonic thickness measurement under the same time section as the benchmark, the linear mapping relationship between the converted values ​​of the other two types of features and the benchmark value is fitted to correct the initial values ​​of the other two types of corrosion rates. After correction, the corrosion rate values ​​corresponding to the three types of features are unified to the same quantization benchmark, completing the dimensional alignment and feature-level fine registration of multi-source features.

[0077] The detection parameters of different dimensions are normalized and uniformly mapped to [0, 1]. The raw and preprocessed data of high-frequency monitoring are stored in a time-series database, and the corrosion assessment results and early warning records of key nodes are stored using blockchain technology.

[0078] After the corrosion assessment report and early warning trigger record of the key node are generated, the on-chain notarization is automatically triggered. First, the core metadata of the data to be notarized is extracted, including the pipe segment number, data generation time, data type, and data version number. At the same time, the hash value of the data is calculated as the data fingerprint. The metadata and data fingerprint are packaged into a notarization transaction and submitted to the corresponding node of the consortium chain. After being verified by a preset number of consensus nodes, it is written into the block and the notarization timestamp is recorded synchronously. After the notarization is completed, a unique notarization certificate number is returned and associated with the data file of the corresponding pipe segment.

[0079] In this embodiment of the invention, the state assessment module constructs a corrosion mechanism feature fingerprint database. Through multi-dimensional feature matching, it automatically identifies and determines four typical corrosion types. For uniform electrochemical corrosion, based on Faraday's law and combined with the corrosion current density obtained by electrochemical impedance spectroscopy, the calculation process simultaneously superimposes the Arrhenius correction term for the temperature of the medium inside the pipe and the accelerated corrosion correction term for the pipe wall stress to obtain the uniform corrosion rate and cumulative corrosion thinning of the pipeline. The remaining wall thickness is obtained by combining the original wall thickness of the pipeline.

[0080] Formula for calculating uniform corrosion rate with temperature-stress double correction:

[0081]

[0082] This represents the corrected uniform corrosion rate of the pipeline, in mm / a.

[0083] The molar mass of the pipe's metal material is expressed in g / mol.

[0084] This represents the corrosion current density detected by the electrochemical impedance spectroscopy sensor, in A / cm².

[0085] This represents the number of electrons gained or lost in the electrochemical reaction of metal corrosion; it is dimensionless.

[0086] denoted by Faraday constant, a physical constant with a value of approximately 96485 C / mol;

[0087] This indicates the density of the pipe's metal material, expressed in g / cm³.

[0088] This represents the apparent activation energy of the corresponding corrosion reaction, expressed in J / mol.

[0089] This represents the ideal gas constant, a physical constant, with a value of approximately 8.314 J / (m²). );

[0090] The reference thermodynamic temperature for representing the corrosion rate is expressed in K.

[0091] This indicates the actual thermodynamic temperature of the medium being transported inside the pipe, expressed in Kelvin (K).

[0092] The stress-accelerated corrosion coefficient is determined by the material and the corrosive environment, and the unit is MPa. -1 ;

[0093] This represents the difference between the actual working stress of the pipe wall and the reference stress, expressed in MPa.

[0094] The fingerprint database pre-inputs characteristic parameter combinations and corresponding judgment conditions for four typical types of corrosion: uniform electrochemical corrosion, local pitting corrosion, under-deposit corrosion, and stray current corrosion. Multiple sets of characteristic thresholds and combination matching rules are set for each type of corrosion. During the evaluation, the wall thickness distribution characteristics, electrochemical impedance spectroscopy characteristics, medium parameter characteristics, and soil environmental parameter characteristics of the current pipe section are extracted and classified into standardized feature vectors according to corrosion type.

[0095] The current feature vector is matched with the corrosion feature templates in the fingerprint database one by one to calculate the similarity. The corrosion type with the highest matching degree and meeting the corresponding judgment threshold is determined as the dominant corrosion type of the current pipe section. If multiple corrosion types meet the matching degree, they are marked as a composite corrosion scenario, and the proportion weight of each corrosion type is recorded simultaneously.

[0096] For local pitting corrosion, based on ultrasonic phased array defect echo imaging, the maximum depth of pitting corrosion, pit area, and local thinning rate are calculated using a pre-calibrated stress-pitting extended finite element simulation inversion model. At the same time, the remaining pressure bearing capacity of the corroded pipe section is calculated by combining the yield strength of the pipe material.

[0097] For under-deposit corrosion, the under-deposit corrosion rate and the degree of interface damage are calculated by combining the scale thickness and interface corrosion characteristics detected by ultrasonic testing and superimposing the under-deposit occlusion cell corrosion acceleration coefficient; for stray current corrosion, the AC corrosion rate and pitting corrosion propagation probability are calculated by combining the coupling relationship between stray current density and corrosion current density.

[0098] Formula for calculating the remaining pressure-bearing capacity of a pipeline with localized pitting corrosion:

[0099]

[0100]

[0101] This indicates the remaining pressure-bearing capacity of the corroded pipe section, expressed in MPa.

[0102] This indicates the yield strength of the pipe material, expressed in MPa.

[0103] This represents the weld coefficient of the pipeline, which is dimensionless and is selected according to the weld type.

[0104] This indicates the original design wall thickness of the pipe, in mm.

[0105] This indicates the maximum depth of pitting defects, in mm.

[0106] This indicates the original outer diameter of the pipe, in mm.

[0107] The bulging factor (Folias factor) is dimensionless.

[0108] Indicates the axial length of pitting defects, in mm;

[0109] This represents the safety factor, which is dimensionless and is selected based on the pipeline's grade and importance.

[0110] The calculation process begins by determining the scale density level based on the scale thickness and interface echo characteristics obtained from ultrasonic testing. Then, a pre-defined table of acceleration coefficients for under-scale occluded cell corrosion is used to select the corresponding acceleration coefficient value. The uniform corrosion rate under the same operating conditions is used as the base value, multiplied by the under-scale corrosion acceleration coefficient to obtain the local under-scale corrosion rate. Finally, the depth and lateral extent of interface corrosion damage are calculated cumulatively based on the scale coverage area and corrosion duration, outputting the under-scale interface damage severity classification result.

[0111] Before inversion, the mechanical properties parameters of the pipe material, such as elastic modulus, Poisson's ratio, and yield strength, as well as the pitting corrosion propagation rate parameters under the corresponding corrosive environment, are imported. Combined with the ultrasonic test sample dataset of standard artificial defects, the finite element simulation model is pre-calibrated and error corrected. During inversion, characteristic parameters such as defect echo amplitude, propagation time, and waveform distortion characteristics obtained from ultrasonic phased array detection are extracted and combined with the current axial stress value of the pipe wall and the medium operating temperature value, and uniformly input into the simulation inversion model.

[0112] The model incorporates multiple pre-set geometric templates for pitting corrosion. By iteratively adjusting geometric parameters such as defect depth, pit diameter, and pit contour, it generates simulated echo features and calculates the deviation between these features and measured echo features until the deviation value falls within a preset convergence threshold. The set of geometric parameters with the smallest deviation is selected as the final inversion result, outputting four core parameters: maximum pitting depth, pit surface area, local wall thickness reduction rate, and equivalent defect diameter.

[0113] Before calculation, the type and geometric dimensions of corrosion defects are determined based on the defect identification results, and the corresponding pressure-bearing capacity calculation model is selected. During calculation, the original outer diameter of the pipeline, the original wall thickness, the minimum remaining wall thickness after corrosion, the depth and axial length of pitting defects, the circumferential width, and the yield strength parameters of the pipeline material are input in sequence. The safety factor and weld coefficient specified in the corresponding standard are selected, and the remaining pressure-bearing capacity is calculated according to the prescribed steps of the relevant standards for pipeline pressure-bearing calculation. The maximum allowable operating pressure value corresponding to the corroded pipe section is output.

[0114] During the identification of under-deposit corrosion, multi-layer echo signals of the pipe wall at full depth are retrieved from an ultrasonic phased array thickness measurement probe. First, gain compensation and noise filtering are performed on the echo signals. Then, multiple interfaces are distinguished based on the acoustic impedance differences between different media interfaces, including the outer surface of the pipe wall, the pipe wall substrate, the scale layer interface, and the inner surface of the scale layer. The thickness of the pipe wall substrate and the scale layer are calculated based on the propagation time difference of the echoes at each interface. At the same time, the echo amplitude, spectrum distribution, and waveform distortion characteristics of the substrate interface below the scale layer are extracted to identify the damage morphology of the under-deposit interface, such as pitting corrosion and uniform corrosion.

[0115] A time-series prediction model integrating corrosion mechanism constraints is constructed. The model takes historical corrosion data, operating parameters, and environmental parameters of the pipeline as inputs, and uses Faraday's law, the Arrhenius temperature effect formula, and the stress-accelerated corrosion formula as physical constraints. The model is embedded in a long short-term memory network model in the form of a loss function regularization term to predict the distribution range of the remaining wall thickness of the pipeline at different future time points. Combined with the pipeline safe wall thickness threshold, the probability range of the remaining service life is calculated. At the same time, the model outputs the changing trend of corrosion development rate and identifies abnormal pipe sections with accelerated corrosion.

[0116] The Long Short-Term Memory (LSTM) network model is configured with an input layer, a first LSTM hidden layer, a second LSTM hidden layer, and a fully connected output layer. The input layer receives sequence feature data arranged at a fixed time step, with the number of nodes matching the dimension of the input features. The input features include six types of parameters: historical corrosion rate, remaining wall thickness, medium temperature, pipe wall stress, soil moisture content, and stray current density. The time step of the input sequence is set to 24.

[0117] The first LSTM hidden layer has 64 neurons and incorporates forget gates, input gates, and output gates. It uses the tanh activation function to process the temporal features of the input sequence. The second LSTM hidden layer has 32 neurons and uses the tanh activation function to further extract the high-dimensional temporal features from the previous layer's output. The number of nodes in the fully connected output layer corresponds to the preset prediction time step, and a linear activation function is used to output the predicted remaining pipe wall thickness at each time step.

[0118] The loss function is composed of a weighted mean square error loss term and a physical regularization term. The mean square error loss term calculates the deviation between the predicted and actual values ​​of the remaining wall thickness. The physical regularization term first calculates the predicted corrosion rate for the corresponding time period by using the difference in the remaining wall thickness between adjacent time steps, and then calculates the theoretical corrosion rate constraint value for the corresponding time step according to Faraday's law, the Arrhenius temperature effect formula, and the stress-accelerated corrosion formula. The deviation between the predicted corrosion rate and the theoretical constraint value is included in the loss calculation. The weight coefficient of the regularization term is set to 0.2.

[0119] The formula for the loss function of a physically constrained time-series prediction model is as follows:

[0120]

[0121]

[0122]

[0123] This represents the total loss function value of the model, and is dimensionless.

[0124] This represents the mean squared error loss term, which measures the numerical deviation between the predicted and actual values ​​of the remaining wall thickness.

[0125] This represents the loss of the physical regularization term, measuring the deviation between the predicted corrosion rate and the theoretical value of the corrosion mechanism.

[0126] This represents the weighting coefficient for the physical regularization term, with a fixed value of 0.2.

[0127] This represents the total number of time steps in the predicted sequence;

[0128] This represents the predicted remaining pipe wall thickness at the i-th time step, in mm.

[0129] This represents the actual remaining pipe wall thickness at the i-th time step, in mm.

[0130] This indicates the actual time length corresponding to a single time step, in years.

[0131] The theoretical corrosion rate at the i-th time step, calculated based on the corrosion mechanism, is expressed in mm / a and is obtained by combining the temperature-stress double-corrected uniform corrosion rate calculation formula with the corresponding operating parameters.

[0132] Before model training, historical monitoring data of the target pipeline network for more than 3 consecutive years were collected to construct a sample dataset, covering pipe sections of different materials and laying areas. Outliers in the dataset were removed using the 3σ criterion, and a sliding time window was used to construct sequence samples with a window length of 24 time steps and a prediction step size of 12 time steps. The processed dataset was randomly divided into training, validation, and test sets in a ratio of 7:2:1. The input features were subjected to min-max normalization to map the data to the interval [0, 1].

[0133] The Adam optimizer was used for iterative model training, with an initial learning rate of 0.001, a training batch size of 32, and a maximum number of iterations of 200. During training, a learning rate decay mechanism was enabled, with the learning rate decreasing to 0.9 times its original value every 20 iterations. An early stopping mechanism was also enabled, terminating training when the validation set loss did not decrease for 10 consecutive iterations, and saving the model weight file with the lowest validation set loss.

[0134] After training, the model accuracy was verified using test set data, with mean absolute error and root mean square error used as accuracy evaluation metrics. During the model inference phase, the corresponding monitoring sequence data was input segment by segment according to the axial mileage of the pipe section, and the remaining wall thickness values ​​for each predicted time step of the corresponding pipe section were output.

[0135] In this embodiment of the invention, the graded early warning module establishes a dynamic risk weight calculation system for the probability of accelerated corrosion evolution. It comprehensively considers static attributes such as pipeline grade, operating pressure, laying area, and surrounding sensitive facilities, combined with the current corrosion type, environmental parameter change trend, and corrosion rate increment. Through the corrosion mechanism model, it calculates the probability of accelerated corrosion of the pipe section in the next 1-3 years, updates the risk weight coefficient of the pipe section, and adjusts the early warning trigger threshold of the pipe section based on the general safe wall thickness threshold and corrosion rate threshold, combined with the dynamic risk weight coefficient. High-risk pipe sections adopt stricter early warning standards. Pipeline grades include main lines, branch lines, and service lines, and laying areas include densely populated areas, industrial areas, and suburbs.

[0136] First, calculate the basic weight coefficient based on the static attributes of the pipe section. Then, calculate the dynamic correction coefficient by combining the current corrosion type, the trend of environmental parameter changes, and the corrosion rate increment. Multiply the basic weight and the dynamic correction coefficient to obtain the final pipe section risk weight coefficient. When adjusting the threshold, the general benchmark threshold is used as the basis. For high-risk pipe sections with a risk weight coefficient greater than 1, the warning trigger threshold is narrowed proportionally according to the weight ratio.

[0137] For low-risk pipe sections with a risk weight coefficient of less than 1, the warning trigger threshold is relaxed proportionally according to the weight ratio; the adjusted warning thresholds for each level are simultaneously stored in the configuration parameter table of the corresponding pipe section, overwriting the original threshold parameters, for subsequent real-time warning determination.

[0138] The system sets up four warning levels: blue, yellow, orange, and red. It also sets three judgment dimensions: corrosion rate level and increment, absolute corrosion thinning, and remaining service life. Meeting the threshold for any of these dimensions triggers a level-four warning. Each warning level is configured with a corresponding quantitative threshold. A blue warning indicates a corrosion rate slightly higher than the industry benchmark under similar operating conditions, with an increasing rate increment and a remaining service life greater than 5 years. A yellow warning indicates local pitting depth exceeding 10% of the wall thickness or a remaining service life of 3-5 years. An orange warning indicates a remaining wall thickness less than 80% of the design wall thickness or a remaining service life of 1-3 years. A red warning indicates a remaining wall thickness approaching the safe wall thickness threshold or the presence of deep pitting with a risk of perforation, and a remaining service life of less than 1 year.

[0139] Before making a judgment, the current dynamic early warning threshold configuration parameters of the corresponding pipe section are retrieved, and the thresholds for each level of the three dimensions of corrosion rate level and increment, absolute corrosion thinning amount, and remaining service life are loaded respectively. The real-time calculated values ​​of the corresponding dimensions of the current pipe section are extracted and compared with the corresponding level thresholds one by one. The early warning level of each dimension is determined independently. If there are no items exceeding the standard, it is judged as a normal state.

[0140] The highest warning level among the three dimensions is taken as the final warning level triggered for this pipeline segment; if multiple dimensions trigger the same warning level simultaneously, the level remains unchanged. After the warning takes effect, a full-dimensional threshold comparison is performed again after each monitoring data update. When the level changes, the warning level is automatically upgraded or downgraded, and a level change record is generated and pushed to the operation and maintenance terminal.

[0141] When an early warning is triggered, the location of the corroded pipe section is automatically located, and the corrosion type, cause analysis, and risk impact range are output simultaneously. The information is pushed to the corresponding maintenance personnel through multiple channels such as system pop-ups, SMS, and mobile APP. At the same time, the early warning level can be adjusted in real time based on monitoring data. The cause analysis includes excessive chloride ions in the soil and excessive dissolved oxygen in the pipe.

[0142] When generating the cause analysis, the soil physicochemical parameters, pipe medium parameters, stray current parameters, and operating parameters of the current pipe section are divided into four dimensions of causes. Within each dimension, the real-time detected values ​​are compared with the preset normal thresholds for corrosion causes one by one, and parameters that exceed the normal range are matched. Corresponding preset corrosion cause tags are associated. When multiple parameters exceed the standard, the cause tags are displayed in order of their weight in terms of their impact on corrosion, and the final cause analysis results are generated.

[0143] In this embodiment of the invention, in the protection effectiveness verification module, for the impressed current cathodic protection device of buried pipeline, when the evaluation finds that the corrosion rate of the outer wall of the pipe section increases and the protection potential deviates from the optimal range, the output current of the corresponding potentiostat is adjusted to stabilize the pipe-to-ground potential relative to the saturated copper sulfate reference electrode in the optimal protection range of -0.85V to -1.10V.

[0144] After adjusting the protection parameters, the electrochemical impedance spectroscopy and power-off potential data of the corresponding pipe section are collected to eliminate the influence of IR drop. The corrosion rate inhibition rate of cathodic protection is quantitatively calculated based on the proportion of the corrosion rate difference before and after the adjustment of the protection parameters. The actual effect of cathodic protection is evaluated, the control parameters are corrected, and the risk of hydrogen embrittlement caused by overprotection is avoided.

[0145] The IR drop elimination method is adopted by instantaneous potential acquisition during power failure. First, the output current of the potentiostat is instantaneously cut off by a remote control command. Simultaneously, the potential acquisition unit is triggered to perform multiple potential acquisitions within a fixed millisecond period after power failure. After three consecutive power failure acquisitions, abnormal acquisition values ​​with deviations exceeding the preset range are eliminated. The average value of the remaining valid acquisition values ​​is taken as the final result. The ohmic voltage drop component generated between the soil medium and the reference electrode is removed to obtain the true polarization potential data of the pipeline.

[0146] The automatic corrosion inhibitor dosing device in the linkage pipe adjusts the concentration and frequency of corrosion inhibitor dosing based on the dissolved oxygen content and corrosion rate of the inner wall. After the dosing parameters are adjusted, the corrosion inhibition efficiency of the corrosion inhibitor is calculated based on the detection data of the corrosion rate of the inner wall. The control logic based on corrosion rate feedback is adopted, and the dosage is gradually reduced when the corrosion rate of the inner wall drops below the target value.

[0147] When reducing the dosage, the corrosion inhibitor concentration is gradually reduced according to a preset fixed ratio, with each reduction not exceeding 20% ​​of the current dosage. Each adjustment of the dosage parameters maintains a preset operating cycle. After the cycle ends, multiple sets of internal wall corrosion rate detection data are continuously collected, and the average value is compared with the target value. If the corrosion rate is still lower than the target value, the next round of reduction is carried out. The dosage is stopped after it drops to the preset minimum maintenance dose. If the corrosion rate rises above the target value during the reduction process, the dosage parameters are adjusted back to the previous level.

[0148] For pipe sections that are low-lying and prone to water accumulation, when the soil moisture content is detected to be continuously exceeding the standard, the surrounding water drainage device is activated. For pipe sections with excessive stray current, the stray current drainage device is activated. After the protective measures are activated, the corrosion rate of the corresponding pipe section is monitored. The protective effect is quantitatively evaluated through the detection data within a preset period. If the expected target is not achieved, the protective parameters are adjusted until the corrosion rate is controlled within the preset range.

[0149] When determining soil moisture content, the soil moisture content detection values ​​of a continuous preset number of sampling periods are statistically analyzed. If all values ​​exceed the preset activation threshold, it is determined to be a continuous exceedance, excluding false triggering caused by single data fluctuations. When determining stray current, the effective average value of stray current density within a unit time window is statistically analyzed. If the average value exceeds the preset activation threshold, the drainage device is triggered to start. After the determination is completed, an activation command is automatically sent to the corresponding protection device, and the command issuance time and the device's feedback operating status are recorded.

[0150] First, the average corrosion rate within a preset period after the protective measures are activated is statistically analyzed and compared with the baseline corrosion rate value of the same period before the activation of the protective measures. The corrosion rate reduction ratio is calculated as the protective effectiveness index. The calculated protective effectiveness index is compared with the preset compliance threshold. If the index reaches or exceeds the threshold, the protective effectiveness is deemed to be up to standard, and the current operating parameters are maintained.

[0151] If the standard is not met, the operating parameters of the corresponding protective device are adjusted upwards according to the preset fixed step size. After the parameter adjustment is completed, wait for the preset stabilization period. After the protective effect is fully effective, collect test data again to evaluate the protective effectiveness. Repeat the parameter adjustment and effectiveness verification process until the protective effectiveness meets the standard or the maximum number of adjustments is reached.

[0152] In this embodiment of the invention, the test and calibration module uses a heat map to display the corrosion level, remaining service life, and warning level of the pipe section. It supports pipe section drilling. Clicking on the pipe section allows users to view real-time sensor data, historical corrosion trends, evaluation reports, warning records, and protection effectiveness verification. It also displays the operating status of monitoring and protection equipment.

[0153] A service life corrosion evolution map and benchmark calibration system were constructed, integrating material corrosion performance test data during pipeline construction, regular offline excavation test data during service, and online continuous monitoring data. Each offline benchmark test result and excavation verification data after emergency response were used as calibration values ​​to iteratively calibrate the parameters and early warning thresholds of the online evaluation model.

[0154] When constructing the map, a two-dimensional coordinate system is established with the pipeline axial mileage as the horizontal axis and the service time as the vertical axis, and the spatial resolution and temporal resolution are set. Data such as corrosion rate, remaining wall thickness, early warning records, and protective measures records of each historical stage are mapped to the corresponding coordinate points of the coordinate system. Spatial interpolation algorithms are used to fill in the parameter values ​​of unmonitored pipe sections to generate a continuous spatiotemporal distribution map of corrosion parameters. The map is divided into multiple independent layers according to parameter type, and supports layer overlay viewing and single layer switching. The map content of the corresponding area is automatically updated after new monitoring data is added.

[0155] During calibration, the time segment and pipe section location corresponding to the offline calibration data are extracted first, and the corresponding output data of the online evaluation model at the same time are retrieved. The offline calibration values ​​and the online model output values ​​are compared group by group, and the deviation values ​​and deviation rates of the three core parameters of corrosion rate, remaining wall thickness and remaining service life are calculated respectively. The correction coefficients and early warning trigger thresholds of the corresponding calculation modules in the model are adjusted proportionally according to the deviation values.

[0156] After the correction is completed, import the historical monitoring dataset of the past year for backtesting and verification to confirm that the deviation rate of the output data after calibration is within the preset allowable range. After the verification is passed, update the model's running parameter library and early warning threshold configuration table, and synchronously record the calibration data source and parameter change records of this calibration.

[0157] It has a built-in hidden danger work order management system. After the warning is triggered, it automatically generates a hidden danger handling work order and dispatches it to the corresponding operation and maintenance team. It tracks the handling progress and results of the work order, supports the formulation of inspection plans and the uploading of inspection records, and integrates manual inspection data with online monitoring data.

[0158] Before merging, the paper or electronic data of manual inspection records are first converted into a structured form, extracting core fields such as corrosion detection values, equipment operating status, and on-site anomaly records, and matching them with the unique number and inspection timestamp of the corresponding pipe section. The converted inspection data is then aligned with the online monitoring data of the same time period and pipe section in terms of time dimension, and the axial mileage of the pipeline is used as the spatial coordinate benchmark to calibrate the spatial positional deviation between the two types of data.

[0159] The parameters of the two types of data are uniformly converted, and the same physical quantities are given uniform units and data precision. The data are then organized into data entries with a uniform format according to a standardized field structure. The aligned data are jointly stored in the historical corrosion database of the corresponding pipe section, and simultaneously incorporated into the corrosion trend analysis dataset and model calibration sample set. The data source is marked as manual inspection or online monitoring.

[0160] The system statistically analyzes the corrosion rate distribution, number of potential hazards, frequency of early warnings, and effectiveness of protective measures for pipelines of different regions, materials, and service life. It aims to uncover the patterns and main causes of corrosion, verify the universality of corrosion mechanism models, and generate periodic reports. The reports include a summary of the current corrosion status, trend analysis, suggestions for hazard mitigation, and operation and maintenance work plans.

[0161] In this embodiment of the invention, when a red alert is triggered and the system assessment determines that there is a risk of leakage or pipe burst, the detection and assessment module activates the encrypted detection and assessment mode. The distributed optical fiber sensing unit is upgraded to a high-frequency continuous sampling mode, the ultrasonic phased array probe is switched to a continuous sampling mode, and the sampling frequency of the electrochemical sensor is encrypted to 1 time / minute. The defect propagation rate and the probability of perforation risk are calculated through the defect propagation inversion model. At the same time, the warning information is pushed to the emergency management department and the operation and maintenance repair team, including the location of the hidden danger, the risk level, the expected impact range, and preliminary handling suggestions.

[0162] After the mode start command is generated, sampling frequency adjustment commands are sent sequentially to the distributed fiber optic sensing unit, ultrasonic phased array thickness probe, and electrochemical impedance spectroscopy sensor according to device priority, setting the high-frequency sampling parameters for the corresponding devices. After the commands are sent, the operating status parameters returned by each device are verified. Once it is confirmed that the device has completed the parameter switching and is operating normally, high-frequency data acquisition and real-time analysis are officially started. During the encrypted mode operation, a device status check is performed at fixed intervals to verify the sampling frequency and data transmission status. An alarm is automatically triggered when a device goes offline or data anomalies occur.

[0163] After the emergency response is completed and the red alert is lifted, a mode recovery command is generated, and a sampling frequency callback command is issued to each device to restore the sampling parameters of each device to the normal operating configuration. After confirming that the device has returned to normal operating status, the encrypted detection and analysis mode is officially exited, and the duration of encrypted mode operation and the amount of data collected are counted simultaneously.

[0164] When calculating the expected impact range, the pipeline topology data and valve layout ledger are retrieved. Combined with the location, diameter, and upstream and downstream valve locations of the potentially hazardous pipe section, a topology traversal algorithm is used to simulate the medium flow path after the potentially hazardous pipe section is cut off. Downstream isolated branch pipe sections and covered users are investigated segment by segment. The total length of affected pipe sections, user types and numbers are counted. At the same time, the geographical coordinate range of the affected area is marked to generate standardized risk impact range data.

[0165] First, the continuously high-frequency collected wall thickness detection data, corrosion rate data, and pipeline operating pressure data are time-series aligned to extract the initial depth, pit area, location information, and other dimensional parameters of the current defect. Then, combined with the fracture toughness parameters and yield strength parameters of the pipeline material and the acceleration coefficient of the corrosion environment, the expansion and change process of the defect size is gradually deduced step by step at fixed small time steps. During the deduction process, the stress alternation effect caused by the operating pressure fluctuation is simultaneously superimposed, and the depth expansion and area expansion of the defect are calculated step by step. The real-time expansion rate of the defect and the probability of perforation within the set prediction time are output.

[0166] The system is linked to the heating network dispatching system to adjust the operating parameters of the corresponding pipe section. When necessary, the system remotely controls the shut-off valve to cut off the medium transport in the potentially hazardous pipe section. At the same time, the system activates the backup pipeline to ensure downstream heating. The system is linked to the geographic information system to retrieve information on underground pipelines, valves, manholes and other facilities around the potentially hazardous pipe section. The system is linked to the emergency command system to simultaneously push monitoring data from the accident site.

[0167] After the emergency response is completed, the system automatically records the detection data of the emergency process, including the development process of the hidden danger, the warning time, the response process, and the response results. Combined with the subsequent excavation and detection results, the corrosion assessment model and the warning threshold are corrected.

[0168] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0169] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A system for monitoring and protecting the corrosion status of heating pipelines, characterized in that, Includes the following modules: The data monitoring module adopts a same-section alignment sensing architecture, deploys fiber optic sensing units and multiple types of detection units, and synchronously collects corrosion-related data of the pipeline body, the transported medium and the buried environment. The processing and storage module performs sequential verification, spatiotemporal coarse registration, denoising, feature-level fine registration and normalization on the detection data to complete the temporal storage and reliable evidence preservation of the data. The condition assessment module constructs a quantitative assessment framework that integrates corrosion mechanism and data-driven approaches. It automatically identifies four typical corrosion types and calculates the corresponding corrosion parameters and remaining pressure bearing capacity. Through a time-series prediction model with embedded physical constraints, it outputs the distribution of the remaining pipe wall thickness and the probability range of the remaining service life. The graded early warning module adjusts the early warning trigger threshold based on the risk weight of the pipeline segment and uses multi-dimensional criteria to trigger graded early warnings. The protection effectiveness verification module links the protection device and adjusts the operating parameters, quantitatively calculates the protection effectiveness based on corrosion detection data, and forms an iterative optimization logic for protection parameters. The test and calibration module builds a visual management and control interface and integrates multi-dimensional data to complete model calibration and operation and maintenance statistical analysis. The detection and analysis module activates an encrypted detection and analysis mode in high-risk scenarios, links with external business systems to support emergency response, and processes data synchronously to correct the assessment model and early warning thresholds.

2. The heating pipeline corrosion status monitoring and protection early warning system according to claim 1, characterized in that, The data monitoring module is equipped with distributed fiber optic sensing units along the outer wall of the pipe to collect axial strain and temperature distribution data of the pipe wall, and to identify stress concentration areas caused by corrosion thinning through strain anomalies. At key nodes of the pipeline, the various detection units are arranged in a cross-sectional alignment manner. The ultrasonic phased array thickness gauge and the electrochemical impedance spectroscopy sensor are arranged at corresponding positions on the outer wall of the same circumferential cross-section to form a pipeline body corrosion detection unit. The ultrasonic phased array thickness gauge detects the remaining wall thickness of the pipe and identifies pitting and ulcer-like corrosion morphology features, while the electrochemical impedance spectroscopy sensor calculates the corrosion current density. Dissolved oxygen sensors and conductivity sensors are arranged at corresponding cross-sections inside the pipe to form a transport medium corrosion detection unit, which collects the dissolved oxygen content and ionic conductivity of the hot water inside the pipe. Soil parameter sensors are arranged at intervals along the pipeline to form a buried environment corrosion detection unit, which collects the physicochemical parameters and stray current density of the buried soil.

3. The heating pipeline corrosion status monitoring and protection early warning system according to claim 2, characterized in that, The processing and storage module adopts a feature-level fine registration algorithm based on corrosion physics priors. First, it constructs data rationality verification rules according to corrosion-related physical laws and removes outliers. Then, it unifies the time reference of multi-source data with different sampling frequencies through a time alignment algorithm. Combined with the pipeline coordinate system and alignment layout rules, it maps the detection data to the pipeline axial mileage and circumferential section coordinates to complete the spatiotemporal coarse alignment. After coarse alignment, corresponding denoising algorithms are applied to the different noise characteristics of the detection data. Then, based on the physical mechanism of corrosion, multiple detection features are uniformly mapped to the physical dimension of corrosion rate to complete feature-level fine registration. Normalization is performed on parameters of different dimensions. The detection data is stored in a time-series database, and blockchain technology is used to store the evaluation results and early warning records.

4. The heating pipeline corrosion status monitoring and protection early warning system according to claim 3, characterized in that, The state assessment module constructs a corrosion mechanism feature fingerprint library, automatically identifies four typical corrosion types through feature matching, and for uniform electrochemical corrosion, it combines the corrosion current density detection results with temperature and stress correction terms to calculate the uniform corrosion rate and cumulative thinning amount, and obtains the remaining wall thickness. For localized pitting corrosion, based on the defect echo imaging of the ultrasonic phased array thickness probe, the pitting corrosion characteristic parameters and the remaining pressure bearing capacity are calculated using a pre-calibrated pitting corrosion propagation simulation inversion model. For under-deposit corrosion, the under-deposit corrosion rate and the degree of interface damage are calculated by combining the scale thickness and interface corrosion characteristics and superimposing the corrosion acceleration coefficient. For stray current corrosion, the AC corrosion rate and the pitting corrosion propagation probability are calculated by combining the coupling relationship between stray current density and corrosion current density.

5. The heating pipeline corrosion status monitoring and protection early warning system according to claim 4, characterized in that, The state assessment module constructs a time-series prediction model that integrates corrosion mechanism constraints. It takes historical pipeline corrosion data, operating parameters, and environmental parameters as inputs, corrosion physical laws as constraints, and embeds a long short-term memory network model in the form of a loss function regularization term to predict the distribution range of the remaining pipeline wall thickness at different time points. It also calculates the probability range of the remaining service life by combining the safe wall thickness threshold, completes the quantitative assessment of the corrosion evolution trend, outputs the corrosion rate change trend, and identifies abnormal pipe sections with accelerated corrosion.

6. The heating pipeline corrosion status monitoring and protection early warning system according to claim 5, characterized in that, The graded early warning module establishes a dynamic risk weight calculation system, which comprehensively considers the static attributes of the pipe section, the current corrosion type, the trend of environmental change, and the corrosion rate increment to calculate the probability of accelerated corrosion of the pipe section and update the risk weight coefficient. The early warning trigger threshold is adjusted in combination with the risk weight of the pipe section, and a more stringent early warning standard is adopted for high-risk pipe sections. The system sets up four warning levels: blue, yellow, orange, and red. The corrosion rate level and increment, absolute corrosion thinning amount, and remaining service life are used as the judgment dimensions. The corresponding warning is triggered when any dimension threshold is met. When the warning is triggered, the corroded pipe section is automatically located, and the corrosion type, cause analysis, and risk range are output. The system supports real-time adjustment of the warning level based on the detection data.

7. A monitoring and early warning system for corrosion status of heating pipelines according to claim 6, characterized in that, The protection effectiveness verification module is designed for impressed current cathodic protection devices. When the corrosion rate on the outer wall of the pipe section increases and the protection potential deviates from the optimal range, the output current of the potentiostat is adjusted to stabilize the pipe-to-ground potential within the optimal protection range. After parameter adjustment, electrochemical impedance spectroscopy and power-off potential data are collected. The corrosion inhibition rate is calculated based on the ratio of the corrosion rate difference before and after adjustment, the protection effectiveness is evaluated, and the control parameters are corrected. The automatic corrosion inhibitor dosing device adjusts the concentration and frequency of corrosion inhibitor dosing based on the dissolved oxygen content and the corrosion rate of the inner wall in the pipe. The corrosion inhibition efficiency is calculated by the percentage decrease in the corrosion rate of the inner wall before and after dosing, and the dosing amount is dynamically adjusted using corrosion rate feedback logic.

8. A monitoring and early warning system for corrosion status of heating pipelines according to claim 7, characterized in that, The protection effectiveness verification module activates the surrounding water drainage device for low-lying pipe sections prone to water accumulation when the soil moisture content continuously exceeds the standard; for pipe sections with excessive stray current, it activates the stray current drainage device; after the protection measures are activated, it monitors the change in the corrosion rate of the corresponding pipe section, quantitatively evaluates the protection effectiveness within a preset period, and adjusts the protection parameters if the protection effect does not meet the preset target until the corrosion rate is controlled within the preset range.

9. A monitoring and early warning system for corrosion status of heating pipelines according to claim 8, characterized in that, The visual control interface of the test and calibration module uses a heat map to display the corrosion degree, remaining service life and warning level of the pipe section. It supports pipe section drilling to view the detection data, corrosion trend, evaluation report and protection effectiveness, and simultaneously displays the operating status of monitoring and protection equipment. Construct a service life corrosion evolution map and calibration system, integrate the detection data from multiple stages, and use offline test and excavation verification data as calibration values ​​to iteratively calibrate the evaluation model parameters and the early warning threshold. It has a built-in work order management system to track the progress of handling, integrate inspection data, statistically analyze corrosion patterns, and generate periodic reports.

10. A monitoring and early warning system for corrosion status of heating pipelines according to claim 9, characterized in that, When the red alert is triggered and a risk of leakage or pipe burst is determined, the detection and judgment module activates the encrypted detection and judgment mode, increases the sampling frequency of the distributed optical fiber sensing unit and the ultrasonic phased array thickness probe, encrypts the sampling frequency of the electrochemical impedance spectroscopy sensor, calculates the defect propagation rate and the probability of perforation risk through the defect propagation inversion model, and pushes the early warning information simultaneously. The coordinated pipeline network dispatching system adjusts operating parameters, and if necessary, cuts off the medium transport in the hazardous pipeline section and activates the backup pipeline, while coordinating with the geographic information system and emergency command system to support the response; After the treatment is completed, the detection data during the process is recorded, and the evaluation model and the early warning threshold are corrected based on the excavation detection results.

Images