Method and system for intelligently sensing corrosion and wire breakage of PCCP (prestressed concrete cylinder pipe) and testing influence of PCCP on structural performance

Abstract

The invention discloses a PCCP (prestressed concrete cylinder pipe) corrosion wire breakage intelligent sensing and structural performance influence testing method, which comprises the following steps: S1, pre-embedding stress-strain distributed optical fibers and DAS sensors, and manufacturing a test tube which can be used for a corrosion wire breakage prototype test; s2, installing the manufactured test tube on an internal pressure testing machine, selecting a prestressed steel wire for corrosion wire breaking test, electrifying by using a voltage-stabilized direct-current power supply with the voltage range of 0-32V to accelerate corrosion wire breaking, connecting a positive electrode of the power supply with a red lead, soaking mortar on the surface of the prestressed steel wire by using a sodium chloride solution with the mass fraction of 5%, after the internal pressure of the test tube is loaded to the internal pressure designed by the test, a direct-current power switch is turned on to carry out corrosion wire breaking, and different prestressed steel wires are selected to repeatedly carry out the operation until sufficient data are collected; and S3, data acquisition and analysis. Compared with an existing method, the corrosion wire breaking signal obtained by the method is real, reliable and comprehensive, and the influence rule of the broken prestressed wire on the other prestressed wires on the periphery is accurately obtained.

Description

Technical Field

[0001] The present invention relates to a method and system for intelligently sensing PCCP pipe corrosion and broken wires and testing the effects thereof on structural performance, belonging to the technical field of water conservancy project detection. Background Art

[0002] Prestressed Concrete Cylinder Pipe (PCCP) is a composite pipe that is widely used in long-distance water supply projects due to its excellent performance. The bearing capacity of PCCP pipe mainly comes from high-strength prestressed steel wires. However, during construction and operation, the prestressed steel wires may break due to hydrogen embrittlement or external corrosion. The accumulation of broken wires leads to a decrease in the bearing capacity of the PCCP pipeline, which in turn causes pipe bursts and causes loss of life and property. Therefore, during pipeline operation, it is necessary to monitor PCCP pipeline wire breakage events and understand the health status of the PCCP pipeline after wire breakage occurs. However, it is still difficult to obtain real corrosion wire breakage signals and monitor the health status of PCCP pipelines through experiments.

[0003] Distributed fiber optic sensing technology is a practical and promising monitoring method, offering numerous advantages, including small size, light weight, corrosion resistance, and immunity to electromagnetic interference. It can be embedded within structures or attached to surfaces, making it particularly suitable for long-distance, large-scale monitoring and highly compatible with PCCP pipeline safety monitoring. Distributed fiber optic monitoring systems can collect stress, strain, temperature, and acoustic signals generated by prestressed steel wire breakage in various layers of a PCCP pipeline and analyze them using artificial intelligence algorithms. In recent years, distributed fiber acoustic sensing (DAS) technology has begun to be applied to PCCP pipe breakage monitoring. Using an entire section of optical fiber as a sensor, DAS technology enables highly sensitive detection and precise location of subtle disturbances at any location along the fiber over tens of kilometers. Compared to traditional monitoring technologies, DAS offers advantages such as continuous distribution, long monitoring distance, low signal distortion, ease of large-scale networking, and strong environmental adaptability. Furthermore, the sensor cable can be laid underwater, requiring only single-ended access to the demodulation instrument, facilitating on-site installation and deployment. Combined with pattern recognition algorithms, this system can reliably provide early warning of PCCP pipe sections at risk of bursting. Therefore, this technology is capable of performing monitoring tasks for both newly built and in-service pipelines. However, due to the lack of sufficient PCCP pipe breakage signal data in China, it is still difficult to establish a mature and reliable domestic PCCP pipe breakage monitoring system.

[0004] In the field of PCCP pipe wire breakage testing, some researchers have conducted laboratory corrosion wire breakage tests by applying a corrosive solution to the mortar protective layer and using seepage to corrode the prestressed steel wire. While this method can realistically simulate natural corrosion conditions, the testing process takes up to a year, and there are problems such as the inability to determine the location of the broken wire and difficulty in quantifying the corrosion rate. Patent number CN 109187325B proposes a method for simulating wire breakage due to corrosion. The method involves slotting the protective layer mortar and placing a perforated container. A solution is injected into the container, and the prestressed steel wire and an external electrode are immersed in the solution to accelerate corrosion. While this method can control the corrosion rate and wire breakage location, it removes the mortar on the surface of the prestressed steel wire, causing the prestressed steel wire to lose the mortar's restraining effect, resulting in a difference between the acoustic signal of the broken wire and the actual signal. The method of cutting the broken wire not only requires destroying the mortar protective layer, but also generates a large amount of mechanical noise signals and is only applicable to hydrogen embrittlement wire breakage signals. Existing wire-breaking corrosion tests that don't damage the mortar rely on corrosion solution penetration to break the wire. This results in an excessively slow corrosion rate, often requiring test cycles exceeding six months. Furthermore, both the corrosion area and the corrosion rate are difficult to quantitatively control. Existing wire-breaking corrosion tests that use electrically accelerated corrosion require destroying the mortar to expose the prestressed steel wire, then injecting electrolyte into the cut to perform the accelerated corrosion. This damages the protective mortar layer, preventing reliable wire-breaking signals from being obtained. Summary of the Invention

[0005] Purpose of the invention: In order to overcome the deficiencies in the prior art, the present invention provides a method and system for intelligently sensing broken wire corrosion in PCCP pipes and testing their impact on structural performance. The resulting broken wire corrosion signal is more reliable and comprehensive than that of existing methods, and the influence of broken prestressed steel wires on the remaining surrounding prestressed steel wires can be accurately determined.

[0006] Technical solution: To solve the above technical problems, the present invention provides a method for intelligently sensing PCCP pipe corrosion and broken wires and testing their impact on structural performance, including the following steps:

[0007] S1: Several test tubes were fabricated for corrosion wire breakage prototype testing. Fiber Bragg grating (FBG) strain sensors and fiber Bragg grating (FBG) stress sensors were wrapped around the surfaces of the PCCP tube core concrete, steel cylinder, prestressed steel wire, and mortar, respectively. DAS sensors were installed on the outer surface of the mortar and the inner surface of the tube core concrete. Fiber Bragg grating (FBG) strain sensors, fiber Bragg grating (FBG) stress sensors, and DAS sensors were led out via optical cables. Copper bars were inserted into the mortar as built-in cathodes and led out with black wires. Red wires were also connected to each prestressed steel wire and led out. Mortar spraying was then carried out. After the PCCP tube was fabricated, the optical cables and all lead wires were integrated and connected to the control room.

[0008] S2: Install the fabricated test tube onto the internal pressure testing machine. Select the prestressed steel wire for the corrosion wire breakage test. Power on a regulated DC power supply with a voltage range of 0 to 32V to accelerate the corrosion wire breakage. Connect the positive pole of the power supply to the red wire and the negative pole to the black wire. Saturate the mortar on the surface of the prestressed steel wire with a 5% mass fraction sodium chloride solution. After the internal pressure of the test tube is loaded to the test design internal pressure, turn on the DC power supply to cause corrosion wire breakage and collect test data.

[0009] S3: Data acquisition and analysis: Multiple tests were conducted using different output currents, i.e., corrosion rates, to investigate wire breakage. The degree of mortar cracking and the relationship between the performance changes of each layer and the corrosion rate were observed. The DAS acoustic wave system was used to collect and locate the acoustic wave signals of PCCP pipe wire breakage. Stress and strain sensors were used to collect stress and strain data of each layer of structure surrounding the PCCP pipe wire breakage during the prestressed steel wire corrosion process. This allowed for a quantitative assessment of the impact of PCCP pipe wire breakage on surrounding structures.

[0010] Preferably, step S2 comprises the following steps:

[0011] Step 21: Lay a plastic film on the outside of the PCCP pipe, and use a hose to inject a 5% mass fraction sodium chloride solution into the plastic film to reduce the loop resistance and depassivate the prestressed steel wire. After the broken wire is completely soaked, the wire is cut. Connect the cathode and the sodium chloride solution at both ends of the DC power supply. Turn on the DC power supply for steady current operation, forcing a direct potential difference between the embedded cathode and the prestressed steel wire to accelerate corrosion. Use a data recording device to record current data in real time for corrosion calculation. Determine whether the wire is broken based on the current data and DAS sensor data, and record the data in real time.

[0012] Step 24: Use steel plugs to seal both ends of the PCCP pipe, apply internal pressure by injecting water into the pipe, and gradually increase the internal pressure by 0.1 MPa per level until it is overloaded by 50%. Record data in real time. Stop loading after the internal pressure stabilizes. After the internal pressure stabilizes, perform corrosion and wire breaking. Use a hose to inject 5% mass fraction sodium chloride solution into the plastic film to reduce the loop resistance and depassivate the prestressed steel wire. After the broken wire is completely soaked, perform wire breaking. Turn on the DC power supply for steady current operation to force a potential difference between the embedded cathode and the prestressed steel wire to accelerate corrosion. Use a data recording device to record current data in real time for corrosion calculation. Determine whether wire breaking occurs based on the current data and DAS sensor data, and record data in real time.

[0013] Step 25: Adjust the output current and repeat steps 23 and 24 to obtain multiple sets of test data.

[0014] Preferably, step S3 comprises the following steps:

[0015] S3.1 Real-time collection of monitoring data

[0016] S3.11 collects distributed stress and strain data. Fiber optic sensors embedded in different layers of the PCCP pipe structure collect real-time stress-strain evolution data at various locations throughout the corrosion and wire breakage process based on the Rayleigh or Brillouin scattering principle.

[0017] The S3.12 uses DAS sensors, which form a long baseline array by laying optical cables on the mortar and concrete surface of the core tube. This array monitors the high-frequency acoustic emission signals generated during the wire breakage process. By combining the arrival time difference of the acoustic wave signals with the signal intensity distribution, the broken wire can be accurately located and identified.

[0018] S3.13 Mark the wire number and position parameters of the broken wire;

[0019] S3.2 Data processing and feature extraction

[0020] S3.21. Wire break event extraction and calibration. During corrosion wire breakage testing, when a sharp mutation or a specific frequency signal characteristic occurs in the DAS acoustic wave signal simultaneously with a stress-strain mutation, it is considered a wire breakage event. The system automatically calibrates the wire breakage time and spatial location, establishing a mapping with the corresponding wire number.

[0021] S3.22. Determine the quantitative index of damage impact, construct the output characteristic variable, and calculate the relative ratio of the strain amplitude increased by the prestressed steel wire due to wire breakage, which is defined as the stress transfer factor λ s , the calculation formula is as follows:

[0022]

[0023] Where, ε after and ε before Respectively represent the strain values ​​of the fiber Bragg grating strain sensor before and after the wire is broken;

[0024] The maximum radius range of the strain disturbance caused by the broken wire is calculated, that is, the maximum distance from the position of the broken wire where the strain change rate exceeds the threshold η, which is defined as the damage diffusion radius R d , the calculation formula is as follows:

[0025]

[0026] Where, ε(r,t after ) is the strain value after the wire is broken at a distance r from the broken wire position; ε(r,t before ) is the strain value before the wire is broken at a distance r from the broken wire position; η is the strain change rate threshold, usually 10% to 20%; δ is a small constant to avoid division by zero, usually 10 -6 ; r is the Euclidean distance from the broken wire position;

[0027] The ratio of the remaining load-bearing capacity of the structure after wire breakage to the intact state is calculated and defined as the residual load-bearing capacity ratio R c , the calculation formula is as follows:

[0028]

[0029] Where p fail is the ultimate bearing pressure value after wire breakage, obtained through pressure test or by finite element simulation or degradation model reasoning; design is the design pressure value;

[0030] S3.3 Deep Learning Model Training and Prediction

[0031] S3.31. Dataset division and standardization. Assuming that a wire break occurs at time t, the input samples ε1, ε2, ε3, ε4, and DAS constructed using monitoring data will be DATA and output samples λ s1 ,λ s2 ,λ s3 ,λ s4 、R d5 、R d6 、R d7 、R d8 、R c The test data are standardized. ε1, ε2, ε3, and ε4 represent the strain values ​​of the fiber Bragg grating strain sensors wrapped around the surfaces of the core concrete, steel cylinder, prestressed steel wire, and mortar from tp to t+n, respectively. n and p are integers that can be determined according to the test time and data acquisition frequency. DAS DATA Indicates the value of the DAS sensor before and after the wire is broken, used for wire breakage positioning, λ s1 ,λ s2 ,λ s3 ,λ s4 R represents the stress transfer factor of the core concrete, steel cylinder, prestressed steel wire and mortar. d5 、R d6 、R d7 、R d8 Represents the damage diffusion radius of the core concrete, steel cylinder, prestressed steel wire and mortar; divided into training set, validation set and test set in a ratio of 70%:15%:15%;

[0032] S3.32. Determine the multi-objective loss function: Where α is the loss weighting factor, i = 1, 2, 3, 4, j = 5, 6, 7, 8, MSE is the average of the squares of the differences between the predicted value and the actual value, the predicted value is the output value of the multimodal fusion network CNN-GRU, and the actual value is the value measured by the sensor;

[0033] S3.33. Use the multimodal fusion network CNN-GRU for modeling. The model mainly includes an input layer, a convolutional layer, a flattening layer, a GRU layer, and an output layer. First, the input data enters the CNN. The CNN extracts local features of the data through convolution and pooling operations to generate a feature map. The feature map output by the CNN is flattened into a one-dimensional vector. The flattened vector is input to the GRU. The GRU processes the sequence data through a gating mechanism and learns time dependencies. Finally, through the fully connected layer and activation function, when the loss function value is less than the threshold or the number of cycles reaches the set value, the prediction result λ is output. s1 ,λ s2 ,λ s3 ,λ s4 、R d5 、R d6 、R d7 、R d8 、R c ;

[0034] S3.34. The trained network can be used to intelligently and quantitatively evaluate the impact of PCCP pipe breakage on surrounding structures. That is, by inputting new wire breakage monitoring data into the model, the results of the impact on structural performance can be output in real time.

[0035] Preferably, the S3.32 is obtained by:

[0036] (1) When the number of cycles is less than 100, the loss function is weighted and integrated using a preset constant weighting factor, with values ​​of α1-α4 being 1.0, α5-α8 being 0.5, and α9 being 1.5;

[0037] (2) When the number of cycles is not less than 100, a dynamic loss weighting mechanism is introduced during the training process. Based on the loss value change rate, the relative growth rate ω of each loss item in the mth round of training is calculated. p (m) , p is 1-9, and the weighting factor is determined dynamically from this: Where, It is the loss value of the p-th task during the m-1th training iteration. It is applicable to complex engineering problems where the indicators change dynamically and the prediction target weights need to be continuously adjusted.

[0038] A system for intelligently sensing broken wires in PCCP pipes and testing their impact on structural performance includes: fiber Bragg grating (FBG) strain sensors and fiber Bragg grating (FBG) stress sensors wrapped around the surfaces of the PCCP pipe's core concrete, steel cylinder, prestressed steel wire, and mortar, respectively; DAS sensors connected in series to the mortar surface and the inner surface of the core concrete; a red wire connected at one end to the prestressed steel wire and at the other to the anode of a power supply; a black wire connected at one end to a copper bar and at the other to the cathode of a power supply; and the fiber Bragg grating (FBG) strain sensor, fiber Bragg grating (FBG) stress sensor, and DAS sensor connected to control room signals.

[0039] In this method, a copper strip is attached 1 cm to the axial side of each prestressed steel wire in the PCCP pipeline, acting as a built-in cathode. Black wire leads are connected to the prestressed steel wires with red wire leads. The lead leads are used to connect to an external DC power supply, and the prestressed steel wires are numbered from the center to the sides. The mortar surface is covered with a water-absorbent material such as cotton sponge. A water-retaining material is applied over the water-retaining material, and an injection hose is sealed with an impermeable plastic film over the outer surface of the mortar-saturated material on the pipe wall. A 5% sodium chloride solution is continuously injected through the hose to maintain a moist, conductive state in the corroded area, allowing percolation to penetrate the prestressed steel wire layer. The positive electrode of the DC power supply is connected to the red wire, and the negative electrode is connected to the black wire. The DC power supply is a regulated DC power supply with a voltage range of 0V-32V, capable of operating in adjustable constant voltage and constant current modes, and equipped with data storage capabilities. This records current or voltage changes during power-on for calculation of steel wire corrosion, which is calculated according to Faraday's law of electrolysis.

[0040] The principle of the electric accelerated corrosion device is as follows:

[0041] Electrode reaction:

[0042] Anode (Iron, Fe): Oxidation reaction occurs (reaction formula: Fe-2e-=Fe 2+ ).

[0043] Iron, as an active metal, loses electrons first and is oxidized to ferrous ions (Fe 2+ ).

[0044] Cathode (copper, Cu): reduction reaction occurs (reaction formula: 2H + +2e-=H2↑).

[0045] At the cathode, H + (or H2O) is reduced to hydrogen (H2) and OH- is generated at the same time. Copper only serves as a conductive electrode and does not participate in the reaction.

[0046] Overall reaction: Fe + 2H2O → Fe(OH)2↓ + H2↑

[0047] Fe2+ With OH in solution - Combined to form a white flocculent precipitate Fe(OH)2 (ferrous hydroxide). Fe(OH)2 is easily oxidized to Fe(OH)3 (reddish-brown precipitate) in the air: 4Fe(OH)2+O 2 +2H2O=4Fe(OH)3

[0048] Fe(OH)3 will further decompose into Fe2O3 and Fe3O4 in the air.

[0049] This paper designs a PCCP pipe corrosion wire breakage test data monitoring system. The system consists of DAS sensors, fiber Bragg grating (FBG) strain sensors, and fiber Bragg grating (FBG) stress sensors. Different sensor types are embedded according to the structural characteristics of each PCCP pipe layer. Five circumferential measurement lines are laid out in each section of the pipe, along with three FBG patch strain gauges, aligned with the distributed circumferential orientation of the nondestructive reference pipe. Four longitudinal measurement lines are located at 0°, 90°, 180°, and 270°, forming separate test circuits.

[0050] For the steel cylinder layer, a carbon fiber composite-based optical cable is laid at each circumferential monitoring section. The cables at each section are connected end-to-end to form a single unit. The carbon fiber composite-based optical cable utilizes a dual-core design. In addition to a distributed optical cable measurement line, each monitoring section is equipped with four fiber Bragg grating (FBG) patch strain gauges, arranged symmetrically around the tube (at 0°, 90°, 180°, and 270°).

[0051] For the inner core concrete, a glass fiber composite-based optical cable is laid at each circumferential monitoring section, with the cables at each section connected end to end to form a single unit. Each monitoring section also features a distributed optical cable measurement line and four fiber Bragg grating (FBG) patch strain gauges, symmetrically arranged around the pipe at 0°, 90°, 180°, and 270°.

[0052] The patch strain gauge is a partitioned patch strain gauge with a test range of up to 3000 με. The strain gauge is fixed to the tube body by epoxy bonding, which is tightly combined and has good strain transmission performance.

[0053] The circumferential temperature compensation of the concrete layer in the pipe core adopts a fiber Bragg grating thermometer (probe type), and its test range can reach 200 ° C. The fiber Bragg grating thermometer (probe type) is fixed in the pipe with a cable tie, and is poured together with the concrete in the pipe, fixed in the structure and well coupled with the structure.

[0054] A 2.0mm micro-fixed-point distributed strain sensing optical cable is laid on each longitudinal monitoring section, and the optical cables of each section are connected end to end. The micro-fixed-point distributed strain sensing optical cable is encapsulated and protected by high-wrapping materials such as HY material. Not only is the strength and surface friction resistance improved, making it easy to groove and implant, but the overall stiffness is also greatly reduced. It has good coupling with loose bodies and can be directly buried in the soil to monitor soil deformation.

[0055] For the outer core concrete, a 0.9mm tight sheath strain sensing optical cable and a four-point string 2mm micro fixed-point grating strain sensing optical cable are laid on the circumferential monitoring section. The optical cables of each section are connected end to end to form a whole.

[0056] The tight sheathed strain sensing optical cable and the high-transmission tight sheathed strain sensing optical cable are encapsulated and protected by high-wrapping materials such as HY material. Not only are the strength and surface friction resistance improved, making it convenient for groove carving and implantation, but the overall stiffness is also greatly reduced. It has good coupling with loose bodies and can be directly buried in the soil to monitor soil deformation.

[0057] The micro fixed-point grating strain sensing optical cable is encapsulated and protected by high-wrapping materials such as HY material. Not only is the strength and surface friction resistance improved, making it easy to groove and implant, but the overall stiffness is also greatly reduced. It has good coupling with loose bodies and can be directly buried in the soil to monitor soil deformation.

[0058] For prestressed steel wires, several monitoring sections are selected along the longitudinal axis of the pipe. Distributed bare fiber is laid out in five circumferential directions. When conducting multiple broken wire point tests, the number of points can be increased as appropriate. Four fiber Bragg gratings are laid out on each monitoring section, at 0°, 90°, 180°, and 270°.

[0059] Prestressed steel wire strain distribution monitoring utilizes fiber grating (FBG) sensors attached to the wire surface. Fiber grating (FBG) sensors are unencapsulated and can be directly attached to the surface of the structure for strain monitoring. FBG sensors offer ease of operation, high sensitivity, waterproofness, corrosion resistance, and long-term stability.

[0060] For the protective layer mortar, a glass fiber composite-based optical cable is laid on each circumferential monitoring section, and the optical cables of each section are connected end to end to form a whole. The glass fiber composite-based optical cable adopts a dual-core optical cable design. While laying a distributed optical cable measurement line for each monitoring section, four-point fiber grating patch strain gauges are laid out and installed. The sensors are arranged in a cross-symmetrical pattern around the pipe body (located at 0°, 90°, 180°, and 270°). A 2.0mm micro-fixed-point distributed strain sensing optical cable is laid on each longitudinal monitoring section, and the optical cables of each section are connected end to end to form a whole. Considering that the OFDR test length is within 100m, four measurement lines can be laid on the surface of the mortar protective layer.

[0061] The glass fiber composite-based optical cable uses epoxy resin glue to glue and fix the distributed glass fiber composite-based optical cable to achieve structural strain testing. The attachment matrix of the optical cable is a fabric strip made of materials such as glass fiber cloth and aramid cloth, which is attached to the composite fabric by sewing or weaving. The optical cable has a certain width, which can obtain more contact area to improve coupling and deformation transfer. It is easy to adhere to the concrete surface. At the same time, it is light in weight, soft in texture, simple to lay and not easy to fall off, making the sensor work more reliable. The upper and lower parts of the optical cable are protected by high-strength engineering fabrics, which improves the survival rate of the sensing optical fiber. The elastic modulus of the glass fiber is equivalent to that of concrete, and it can deform synergistically with the concrete pipe body.

[0062] The patch strain gauge is a partitioned patch strain gauge with a test range of up to 3000 με. The strain gauge is fixed to the tube body by epoxy bonding, which is tightly combined and has good strain transmission performance.

[0063] This 2.0mm micro-point distributed strain sensing optical cable is encapsulated and protected with highly resistant materials such as HY material. This not only improves strength and surface friction, facilitating groove insertion, but also significantly reduces overall stiffness. It offers excellent coupling with loose materials and can be directly buried in the soil to monitor soil deformation. It is suitable for surface monitoring of concrete structures and deformation monitoring of model test bodies.

[0064] After all fiber grating sensors and distributed optical cables are laid out, they are introduced into the established monitoring station using optical fiber leads and optical cables, so that all optical path signals can collect data in the monitoring station and realize centralized monitoring of the system.

[0065] The patch strain gauges and thermometers of the inner core concrete have four pieces welded in series in each section, with lead wires leading out, occupying a total of 9 channels. The patch strain gauges of the steel cylinder have four pieces welded in series in each section, with lead wires leading out, occupying a total of 3 channels. The micro fixed-point grating strain sensing optical cable of the outer core concrete has four points in each section, with lead wires leading out, occupying a total of 3 channels. The high-strength fiber grating points of the prestressed steel wire have four pieces welded in series in each section, with lead wires leading out, occupying a total of 9 channels. The patch strain gauges of the mortar protective layer have four pieces welded in series in each section, with lead wires leading out, occupying a total of 9 channels. After the fiber grating sensors at the above five layer positions are welded in series, they occupy a total of no less than 24 channels of the demodulator.

[0066] Each layer of the distributed optical cable is connected to a demodulator as a separate loop for testing. The ends of the glass fiber composite-based optical cable in the inner concrete tube are individually led out of the hole. The ends of the steel tube carbon fiber composite-based optical cable are individually led out of the hole. Each circumferential optical cable in the outer concrete tube carbon fiber composite-based optical cable is independently led out of the mortar protective layer. After the structure is completed, the circumferential sections are connected in series to form a complete test line. The ends of the glass fiber composite-based optical cable in the mortar protective layer are individually led out of the hole. Sixteen DAS sensing optical fibers are attached to the outer surface of the protective mortar with impregnated adhesive. A total of 16 fibers are arranged, evenly distributed circumferentially. All locations where the fiber grating sensor lines and distributed optical cable lines overlap are uniformly led out of the hole.

[0067] In the present invention, a copper cathode is pre-buried when producing the test tube, and the prestressed steel wire is pre-connected with a conductor, so that the mortar protective layer can be avoided from being damaged in the subsequent test process. The purpose of the salt solution penetration in the present invention is to construct a complete electrolytic cell reaction when using the salt solution for penetration rather than as a corrosion solution. (After penetration, a current loop is formed: positive electrode of power supply > prestressed steel wire > mortar pore liquid > pre-buried cathode > negative electrode of power supply) The advantages of this method over the existing method of splitting the mortar protective layer and using electric accelerated corrosion test are: 1. The constraint of the mortar protective layer on the prestressed steel wire above the broken wire is retained, while splitting the mortar will change the actual vibration mode of the prestressed steel wire after the wire is broken, resulting in the collected acoustic wave signal being inconsistent with the actual situation. 2. The corrosion range is more concentrated, the current loop is more concentrated and narrow, the reaction rate is high, and the position of the broken wire is determined by the position of the buried cathode. 3. The electrolyte is in the form of mortar pore liquid, which will not cause the corrosion products to be carried out by the electrolyte. It can restore the effect of rust expansion on the mortar during the corrosion of the prestressed steel wire and restore the mortar cracking process under real corrosion and broken wire conditions. Regarding the location of broken wires: Each prestressed steel wire has a pre-buried cathode corresponding to it (i.e., a prestressed steel wire is selected for breaking by connecting it to the adjacent cathode through a DC power supply). The corrosion range is concentrated at the position directly opposite the copper electrode (there is a certain diffusion angle, which can be further controlled by applying a bonding coating to the prestressed wire during production). The corrosion rate can be controlled by the power supply (the corrosion rate is linearly related to the current).

[0068] This technical solution utilizes a DAS acoustic system to collect and locate acoustic signals from PCCP pipe breakage caused by corrosion. It also utilizes a stress-strain distributed optical fiber system to collect stress and strain data from various layers of structures surrounding the PCCP breakage during the corrosion process. This allows for a quantitative assessment of the impact of PCCP breakage on surrounding structures. This technical solution uses experimental data to establish and train a model for practical application in real-world projects. In real-world projects, breakages do not occur at a single point; rather, they occur simultaneously or sequentially at multiple points. Breakages are often distributed randomly across multiple points. Multiple breakages result in highly nonlinear stress redistribution paths, with the effects of each breakage point superimposed and interfering with each other, forming a complex strain field. The strain signals generated by multiple breakages can "couple" and interfere with each other within the structure. This is especially true when distributed optical fiber and DAS capture spatially continuous data. The breakage signals overlap and blur, making it difficult to clearly identify their boundaries. The strain transfer zones may completely overlap, making it nearly impossible to infer the independent impact of each breakage point using traditional mechanical models. Therefore, it is crucial to develop a deep learning model that integrates multivariate fusion and nonlinear reasoning capabilities.

[0069] In this technical solution, in step S3, data of each working condition is collected, and the stress and deformation state of each layer of the pipeline structure is collected regularly to analyze the changes in the stress and deformation state of each layer of the pipeline structure without external load and corrosion and wire breakage; the stress and deformation of each layer of the test pipe without corrosion process under internal pressure is analyzed; under no internal pressure working condition, after the wire is broken by power corrosion, the distributed optical fiber and each measuring point data are collected in real time until the wire break signal occurs and stabilizes, and the influence of the no internal pressure corrosion and wire break process on the stress and deformation of each layer of the test pipe structure is analyzed; under the operating internal pressure working condition, after the wire is broken by power corrosion, the distributed optical fiber and each measuring point data are collected in real time. The optical fiber and the data of each measuring point are collected until the broken wire signal occurs and stabilizes, and the influence of the internal pressure corrosion and broken wire process on the stress and deformation of each layer structure of the test tube is analyzed; the working conditions without internal pressure are compared with the working conditions with internal pressure during operation, and the influence of the presence or absence of internal pressure on the stress and deformation of each layer structure of the test tube and the influence on the broken wire signal under the action of corrosion and broken wire is analyzed; the prototype test of broken wire without internal pressure and the prototype test of broken wire with designed internal pressure are compared with different output current sizes, and the influence of different corrosion rates on the structure of each layer of the test tube and the broken wire signal is analyzed; the broken wire acoustic wave signal collected by DAS is identified by combining deep learning.

[0070] Beneficial effects: The corrosion broken wire signal acquisition test method of the present invention is innovative. Through non-destructive accelerated corrosion, a more reliable broken wire signal is obtained. The stress and strain data of each layer structure around the broken wire of the PCCP pipe during the prestressed steel wire corrosion process are collected by using a stress-strain distributed optical fiber system to quantitatively evaluate the impact of the PCCP pipe broken wire on the surrounding structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a schematic diagram of the technical route of the present invention.

[0072] Figure 2 Schematic diagram of the PCCP pipe of the present invention.

[0073] Figure 3 Schematic diagram of the prototype test device.

[0074] Figure 4 Schematic diagram of the impact of broken wire.

[0075] Figure 5 Schematic diagram of current path and steel wire corrosion degree.

[0076] Figure 6 Schematic diagram of the layout of stress-strain optical fiber and DAS sensors.

[0077] In the figure: 1. Test tube; 2. Steel bulkhead; 3. Pressurized water pipe; 4. Corrosion-broken wire point; 5. Lead wire; 6. Control room; 7. Ground; 11. Mortar surface water-retaining material; 12. Protective layer mortar; 13. Cathode; 14. Pipe core concrete; 15. Red wire; 16. Black wire; 17. Prestressed steel wire; 18. DC power supply. DETAILED DESCRIPTION

[0078] The present invention will be further described below with reference to the accompanying drawings.

[0079] like Figures 1 to 6 As shown, the PCCP pipe of the present invention includes a core concrete, a steel cylinder, prestressed steel wire and a mortar layer. A mortar surface water-retaining material is provided on the outside of the mortar layer. Several copper bars are built into the core concrete as cathodes. Each copper bar is arranged in pairs with each prestressed steel wire. Fiber Bragg grating strain sensors and fiber Bragg grating stress sensors are provided in the core concrete, steel cylinder, prestressed steel wire and mortar layer. DAS sensors (distributed acoustic wave sensors) are connected in series on the mortar surface and the inner surface of the core concrete. A copper bar is inserted into the mortar as a built-in cathode and is led out with a black wire. A temperature sensor is also provided in the PCCP pipe. A red wire is connected to each prestressed steel wire, and a black wire is connected to the cathode. The red wire and the black wire are connected to a DC power supply.

[0080] A method for intelligently sensing PCCP pipe corrosion and broken wires and testing their impact on structural performance includes the following steps:

[0081] S1: A test tube suitable for corrosion wire breakage prototype testing was fabricated. Fiber Bragg grating (FBG) strain sensors and fiber Bragg grating (FBG) stress sensors were implanted on the inner and outer surfaces of the PCCP core concrete, the steel cylinder, the prestressed steel wires, and the mortar surface. Copper strips were attached to the 10 prestressed steel wires in the middle section of the PCCP tube at a 1 cm axial offset, acting as built-in cathodes. These were led out with black wires. A red wire was connected to each prestressed steel wire and led out. Mortar spraying was then performed. After the tube was fabricated, the fiber optic system leads were integrated and connected to the control room.

[0082] Step 11: Distributed fiber optic routing on the steel cylinder: Use a grinder to polish the steel cylinder surface at designated points and lines, with a width of 10 cm, to ensure a smooth and flat surface. Attach the fiber grating (FBG) patch strain gauge and secure it to the cylinder surface using spot welding. Before attaching the FBG patch strain gauge, first apply a non-sticky white silicone coating to the grating area to prevent epoxy from entering and affecting proper operation. For carbon fiber composite (CFRP) cable routing, apply a layer of CFRP impregnating adhesive to the cable routing route to improve adhesion between the carbon fiber and the steel pipe. Next, lay the CFRP cable along the route, ensuring a straight routing process. Pre-tension the cable as it is laid, and use rollers to remove any bubbles between the CFRP cable and the primer. Finally, apply another coat of impregnating adhesive to the entire cable route, ensuring complete penetration of the carbon fiber throughout the route to maximize protection for the sensing cable. Thread the lead wires through a soft rubber tube for protection. A protective box is set up at the outlet position, and a part of the lead wire is coiled in the box for backup.

[0083] Step 12: Distributed fiber optic cable routing in the outer tube concrete. The 0.9mm tight-jacketed strain sensing cable was installed and secured using surface-grooved epoxy potting and then covered with glass cloth. The 2mm micro-fixed-point grating strain sensing cable was installed and secured using epoxy resin adhesive and then covered with glass cloth. The sensor lines were protected by a junction box. A separate junction box was used for temporary protection in each ring. Later, the junction box was opened, the leads were serially spliced, and then connected to the corresponding fiber optic interrogator for testing.

[0084] After each ring-shaped cable is laid, when transitioning to the next ring-shaped cable, the optical cable does not need to be cut. One redundant end is directly fixed on the surface of the structure as a transition section, and the next ring-shaped cable is fixed and installed, completing the laying of each ring-shaped cable in turn.

[0085] When installing the fiber grating patch strain gauge, first grind the concrete surface smooth and flat. After preliminary fixation with paper tape, temporarily fix the two ends with glue, and then apply epoxy resin glue to the entire surface. The outgoing wires at both ends of the sensor are protected by wearing a 3mm loose-tube armored hollow tube. Then, four sensors are connected in series in a circular manner and the lead wires are led out.

[0086] On-site installation requires the installation of fiber Bragg grating (FBG) patch strain gauges first, followed by the installation of distributed optical cables. Glass fiber composite-based optical cables can be directly applied to the surface of the patch strain gauges. All sensors are installed and deployed after the inner core concrete structure is completed and before the test tube is erected. The cables are routed out of the tube through perforations in the tube cover.

[0087] Fiber Bragg grating thermometers (probe type) are secured within the pipe using cable ties and cast alongside the concrete, securing the cable within the structure and achieving good coupling. After each loop is laid, the redundant end of the cable is secured to the next loop without cutting. All sensors are integrated after the inner core concrete structure is completed and before the test pipe is erected. The wiring is routed out of the pipe through perforations in the pipe cover.

[0088] For differential fixed-point distributed strain sensing optical cables, the optical cable laying route is first marked, and the optical cable is pre-tensioned while being laid. After tensioning, it is fixed and the optical cable is fixed inside the PCCP tube with epoxy resin glue. After the first epoxy resin glue solidifies, another layer is added on the original basis to achieve the best protection effect for the sensing optical cable. An independent junction box is used for temporary protection. Later, the junction box is opened and the leads are fused in series and then inserted into the corresponding optical fiber demodulator for testing.

[0089] Step 13: Inner core concrete distributed optical fiber layout. When installing glass fiber composite-based optical cable, first clean the surface of the structure, remove impurities, apply a layer of quick-drying glue on the surface of the structure, then lay the optical cable, use a roller to press the optical cable tightly, and then apply a layer of epoxy resin glue on the upper part of the optical cable, press it tightly with a roller again, and then use cloth-based tape to temporarily fix the optical cable to prevent it from falling off. After the epoxy glue cures for about 24 hours, it can be firmly fixed on the surface of the structure and achieve good coupling with the structure.

[0090] Step 14: Prestressed steel wire distributed optical fiber layout. After the prestressed steel wire is coiled, the location of each monitoring point is determined according to the design requirements. The steel wire at the monitoring point is polished and cleaned with sandpaper. The FBG point is temporarily fixed to the steel wire with paper tape. The two ends are temporarily fixed with glue. The ends of the FBG point are sleeved with 3mm loose-tube armored hollow tubes to protect the leads at both ends. Then, epoxy resin is applied to the entire length of the FBG point. After curing for about 24 hours, the sensor layout is complete.

[0091] To improve the survival rate, each measurement point was placed on the side of the steel wire. The same process was used to lay the distributed strain gauge bare fiber and distributed optical cable. The sensor lines were protected by a junction box, and each ring was temporarily protected by an independent junction box. Later, the junction box was opened, the leads were serially spliced, and then connected to the corresponding fiber optic interrogator for testing.

[0092] Step 15: Lay out the corrosion-wire breaking device. After the prestressed steel wire is wrapped, select the corrosion-wire breaking points and number the prestressed steel wires at the selected points from the center to the sides. Select 10 or 20 steel wires depending on the test conditions. Pre-embed a copper cathode 1 cm from one side of the prestressed steel wire and lead it out with a black leather-sheathed wire. Sand the surface of the prestressed steel wire with sandpaper to remove the passivation film, and lead it out with a red wire. After the protective layer mortar is roller-blasted, apply water-retaining material to the corrosion points, bury a water pipe and cover it with plastic film, and continuously inject water to soak the protective layer mortar. Turn on the power and determine whether the solution has penetrated the prestressed steel wire based on the current. Use a regulated DC power supply with a voltage range of 0V-32V to accelerate corrosion.

[0093] Step 16: Protective mortar distributed fiber optic layout

[0094] For differential fixed-point distributed strain sensing optical cables, the optical cable laying route is first marked, and the optical cable is pre-tensioned while being laid. After tensioning, it is fixed and the optical cable is fixed to the surface of the PCCP pipe mortar protective layer with epoxy resin glue. When the first epoxy resin glue solidifies, another layer is added on the original basis to achieve the best protection effect for the sensing optical cable. An independent junction box is used for temporary protection. Later, the junction box is opened and the leads are fused in series and then inserted into the corresponding optical fiber demodulator for testing.

[0095] Step 17: DAS sensor layout: DAS sensors are placed on the surface of the protective layer mortar, glued with impregnated glue, arranged longitudinally, and 16 sensors are evenly distributed along the circumference.

[0096] Step 18: System Integration: The optical fiber leads from each layer of the PCCP pipe are numbered and centrally connected to the control room and connected to the demodulation equipment. The control room controls the internal pressure loading of the test device and monitors the distributed optical fiber system and acoustic sensor signal feedback in real time.

[0097] S2: Install the fabricated test tube onto the internal pressure testing machine, select the prestressed steel wire for the corrosion wire breakage test, apply power to a regulated DC power supply with a voltage range of 0 to 32V to accelerate the corrosion wire breakage, connect the positive pole of the power supply to the red wire, and the negative pole of the power supply to the black wire, soak the mortar on the surface of the prestressed steel wire with a 5% mass fraction sodium chloride solution, load the internal pressure of the test tube to the test design internal pressure, turn on the DC power supply switch to cause corrosion wire breakage, and repeat the above operation with different prestressed steel wires until sufficient data is collected;

[0098] Step 21: Static Pipe Test: Set up a static test pipe without external loads and record the time-dependent stress and deformation behavior of each layer of the pipe. This serves as a control group to observe the effects of internal pressure alone and corrosion wire breakage alone on the stress and deformation behavior of the PCCP pipe structure.

[0099] Step 22: Prototype test of wire breakage due to corrosion without internal pressure. Carry out wire breakage due to corrosion without internal pressure. Use a hose to inject 5% mass fraction sodium chloride solution into the plastic film until the broken wire is completely soaked. Then break the wire. Turn on the DC power supply for steady current operation. Use a data recording device to record the current data in real time for corrosion calculation. Determine whether the wire breakage occurs based on the current data and DAS sensor data. Record the data in real time. After the signal stabilizes, replace the prestressed steel wire for wire breakage test.

[0100] Step 23: Internal pressure prototype test, install the test tube on the internal pressure testing machine, seal both ends with steel plugs, apply internal pressure by injecting water into the tube, load the internal pressure step by step, load 0.1MPa at each stage, load to the designed internal pressure, record data in real time, stop loading after the internal pressure stabilizes, perform wire breaking by corrosion after the internal pressure stabilizes, use a hose to inject 5% mass fraction sodium chloride solution into the plastic film, and break the wire after it is completely soaked, turn on the DC power supply for steady current operation, use a data recording device to record the current data in real time for corrosion calculation, judge whether the wire is broken based on the current data and DAS sensor data, record data in real time, replace the prestressed steel wire for wire breaking test after the signal stabilizes.

[0101] Step 24: Adjust the output current and repeat the prototype test of no internal pressure corrosion wire breakage and the prototype test of designed internal pressure corrosion wire breakage.

[0102] S3: Data acquisition and analysis: Multiple tests were conducted using different output currents, i.e., corrosion rates, to investigate wire breakage. The degree of mortar cracking and the relationship between the performance changes of each layer and the corrosion rate were observed. A DAS acoustic wave system was used to collect and locate the acoustic wave signals of PCCP pipe wire breakage. A stress-strain distributed optical fiber system was used to collect stress and strain data of each layer of structure surrounding the PCCP pipe wire breakage during the prestressed steel wire corrosion process. This allowed for a quantitative assessment of the impact of PCCP pipe wire breakage on surrounding structures.

[0103] S3.1 Real-time collection of monitoring data

[0104] 1. Collect distributed stress and strain data. Fiber optic sensors embedded in different layers of the PCCP pipe structure (e.g., core concrete layer, steel cylinder, steel wire, and mortar) collect real-time stress-strain evolution data at various locations throughout the corrosion and wire breakage process based on the Rayleigh or Brillouin scattering principle.

[0105] 2. A distributed acoustic sensing system (DAS) uses a long-baseline array of optical cables laid through the steel wire layer to monitor the high-frequency acoustic emission signals generated during wire breakage. Combining the acoustic signal arrival time difference with the signal intensity distribution, the broken wire can be accurately located and identified.

[0106] 3. Mark the position parameters of the structural space information such as the wire number where the broken wire is located, the coordinates of the broken wire position, etc.

[0107] S3.2 Data processing and feature extraction

[0108] 1. Wire break event extraction and calibration. In corrosion wire breakage tests, when the DAS acoustic wave signal shows a sharp mutation or a specific frequency signal feature that occurs synchronously with a stress-strain mutation, it is considered a wire breakage event. The system automatically calibrates the wire breakage time and spatial position. This can be accomplished using existing technology to establish a mapping with the corresponding numbered steel wire.

[0109] 2. Determine the quantitative index of damage impact, construct the output characteristic variable, and calculate the relative ratio of the strain amplitude increased by the prestressed steel wire due to wire breakage, which is defined as the stress transfer factor λ s This indicator can quantify how much the force on the wire changes and whether it is likely to become the "next wire breaking point". The calculation formula is as follows:

[0110]

[0111] Where, ε after and ε before Respectively represent the stress on the adjacent steel wires before and after the wire breaks.

[0112] The maximum radius range of the strain disturbance caused by the broken wire is calculated, that is, the maximum distance from the position of the broken wire where the strain change rate exceeds the threshold η, which is defined as the damage diffusion radius R d This indicator is used to determine the local or global structural safety impact range and provide intervention boundaries for auxiliary decision-making detection and maintenance. The calculation formula is as follows:

[0113]

[0114] Where, ε(r,t after ) is the strain value after the wire is broken at a distance r from the broken wire position; ε(r,t before ) is the strain value before the wire is broken at a distance r from the broken wire position; η is the strain change rate threshold, usually 10% to 20%; δ is a small constant to avoid division by zero, usually 10 -6 r is the Euclidean distance from the broken wire. r represents the distance between the strain measurement point and the broken wire. It can be expressed as an absolute distance in space, i.e., the Euclidean distance, or as the pipe diameter and horizontal angle. The r value is equivalent to the positioning coordinate. Once the broken wire location is determined, it can be pinpointed using strain distributed optical fiber positioning.

[0115] The ratio of the remaining load-bearing capacity of the structure after wire breakage to the intact state is calculated and defined as the residual load-bearing capacity ratio R c , this indicator is used to estimate the remaining life and safety margin, and the calculation formula is as follows:

[0116]

[0117] Where p fail is the ultimate bearing pressure value after wire breakage, which can be obtained through pressure test, finite element simulation or degradation model reasoning; design is the design pressure value.

[0118] S3.3 Deep Learning Model Training and Prediction

[0119] S3.31. Dataset division and standardization. Assuming that a wire break occurs at time t, the input samples ε1, ε2, ε3, ε4, and DAS constructed using monitoring data will be DATA and output samples λ s1 ,λ s2 ,λ s3 ,λ s4 、R d5 、R d6 、R d7 、R d8 、R c The test data are standardized. ε1, ε2, ε3, and ε4 represent the strain values ​​of the fiber Bragg grating strain sensors wrapped around the surfaces of the core concrete, steel cylinder, prestressed steel wire, and mortar from tp to t+n, respectively. n and p are integers that can be determined according to the test time and data acquisition frequency. DAS DATA Indicates the value of the DAS sensor before and after the wire is broken, used for wire breakage positioning, λ s1 ,λ s2 ,λ s3 ,λ s4 R represents the stress transfer factor of the core concrete, steel cylinder, prestressed steel wire and mortar. d5 、R d6 、R d7 、R d8 Represents the damage diffusion radius of the core concrete, steel cylinder, prestressed steel wire and mortar; divided into training set, validation set and test set in a ratio of 70%:15%:15%;

[0120] S3.32. Determine the multi-objective loss function: Where α is the loss weighting factor, i = 1, 2, 3, 4, j = 5, 6, 7, 8, and MSE is the average of the squares of the differences between the predicted and actual values. N is the number of samples, Y K is the predicted value, is the measured value, the predicted value is the output value of the multimodal fusion network CNN-GRU, and the actual value is the value measured by the sensor;

[0121] S3.33. Use the multimodal fusion network CNN-GRU for modeling. The model includes an input layer, a convolutional layer, a flattening layer, a GRU layer, and an output layer. First, the input data enters the CNN. The CNN extracts local features of the data through convolution and pooling operations to generate a feature map. The feature map output by the CNN is flattened into a one-dimensional vector. The flattened vector is input to the GRU. The GRU processes the sequence data through a gating mechanism and learns time dependencies. Finally, through the fully connected layer and activation function, when the loss function value is less than the threshold or the number of cycles reaches the set value, the prediction result λ is output. s1 ,λ s2 ,λ s3 ,λ s4 、R d5 、R d6 、R d7 、R d8 、R c ;

[0122] S3.34. The trained network can be used to intelligently and quantitatively evaluate the impact of PCCP pipe breakage on surrounding structures. By inputting new wire breakage monitoring data into the model, the results of the impact on structural performance can be output in real time.

[0123] The S3.32 is obtained by:

[0124] (1) When the number of cycles is less than 100, the loss function is weighted and integrated using a preset constant weighting factor, with values ​​of α1-α4 being 1.0, α5-α8 being 0.5, and α9 being 1.5;

[0125] (2) When the number of cycles is not less than 100, a dynamic loss weighting mechanism is introduced during the training process. Based on the loss value change rate, the relative growth rate of each loss item in the mth round of training is calculated. p is a natural number from 1 to 9, and the weighting factor is dynamically determined based on it:

[0126]

[0127] Where, The loss value of the p-th task at the m-1th iteration of training. In this way, tasks with slower convergence are given greater weights in the early stages of training, which promotes consistent training progress for all tasks. This is suitable for complex engineering problems where indicators change dynamically and the prediction target weights need to be continuously adjusted.

[0128] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for intelligently sensing PCCP pipe corrosion and broken wires and testing their impact on structural performance, characterized in that: The following steps are involved: S1: Several test tubes were fabricated for prototype corrosion and wire breakage testing. Fiber Bragg grating (FBG) strain sensors and fiber Bragg grating (FBG) stress sensors were wound around the surfaces of the PCCP tube core concrete, steel tube, prestressed steel wire, and mortar, respectively. DAS sensors were installed on the outer surface of the mortar and the inner surface of the tube core concrete. Fiber Bragg grating (FBG) strain sensors, fiber Bragg grating (FBG) stress sensors, and DAS sensors were led out via optical cables. Copper bars were inserted into the mortar as built-in cathodes and led out with black wires. Red wires were also connected to each prestressed steel wire and led out. Mortar spraying was then carried out. After the PCCP tube was fabricated, all leads were integrated and connected to the control room. S2: Install the fabricated test tube onto the internal pressure testing machine. Select the prestressed steel wire for the corrosion wire breakage test. Power on a regulated DC power supply with a voltage range of 0 to 32V to accelerate the corrosion wire breakage. Connect the positive pole of the power supply to the red wire and the negative pole to the black wire. Saturate the mortar on the surface of the prestressed steel wire with a 5% mass fraction sodium chloride solution. After the internal pressure of the test tube is loaded to the test design internal pressure, turn on the DC power supply to cause corrosion wire breakage and collect test data. S3: Data acquisition and analysis. Multiple tests were conducted using different output currents, i.e., corrosion rates, to investigate wire breakage. The degree of mortar cracking and the relationship between the performance changes of each layer and the corrosion rate were observed. The DAS sensor system was used to collect and locate the acoustic wave signals of PCCP pipe wire breakage. Stress and strain sensors were used to collect stress and strain data of each layer of structure surrounding the PCCP pipe breakage during the prestressed steel wire corrosion process. This allowed for a quantitative assessment of the impact of PCCP pipe breakage on surrounding structures.

2. The method for intelligently sensing broken wires in PCCP pipes and testing their impact on structural performance according to claim 1 is characterized in that: The step S2 comprises the following steps: Step 21: Lay a plastic film on the outside of the PCCP pipe, and use a hose to inject a 5% mass fraction sodium chloride solution into the plastic film to reduce the loop resistance and depassivate the prestressed steel wire until the broken wire is completely soaked. Connect the cathode and the sodium chloride solution at both ends of the DC power supply, turn on the DC power supply for steady current operation, and force a direct potential difference between the embedded cathode and the prestressed steel wire to accelerate corrosion. Use a data recording device to record current data in real time for corrosion calculation. Determine whether the wire is broken based on the current data and DAS sensor data, and record the data in real time. Step 24: Use steel plugs to seal both ends of the PCCP pipe. Apply internal pressure by injecting water into the pipe. The internal pressure is loaded step by step, with each level loading 0.1 MPa, until the test design internal pressure is reached. Record data in real time. Stop loading after the internal pressure stabilizes. After the internal pressure stabilizes, perform corrosion and wire breaking. Use a hose to inject 5% mass fraction sodium chloride solution into the plastic film to reduce the loop resistance and depassivate the prestressed steel wire. After the broken wire is completely soaked, perform wire breaking. Turn on the DC power supply for steady current operation to force a potential difference between the embedded cathode and the prestressed steel wire to accelerate corrosion. Use a data recording device to record current data in real time for corrosion calculation. Determine whether wire breaking occurs based on the current data and DAS sensor data, and record data in real time. Step 25: Adjust the output current and repeat steps 23 and 24 to obtain multiple sets of test data.

3. The method for intelligently sensing broken wires in PCCP pipes and testing their impact on structural performance according to claim 2 is characterized in that: The step S3 comprises the following steps: S3.1 Real-time collection of monitoring data S3.11 collects distributed stress and strain data. Fiber optic sensors embedded in different layers of the PCCP pipe structure collect real-time stress-strain evolution data at various locations throughout the corrosion and wire breakage process based on the Rayleigh or Brillouin scattering principle. The S3.12 uses DAS sensors, which form a long baseline array by laying optical cables on the mortar and concrete surface of the core tube. This array monitors the high-frequency acoustic emission signals generated during the wire breakage process. By combining the arrival time difference of the acoustic wave signals with the signal intensity distribution, the broken wire can be accurately located and identified. S3.13 Mark the wire number and position parameters of the broken wire; S3.2 Data processing and feature extraction S3.

21. Wire break event extraction and calibration. During corrosion wire breakage testing, when the DAS sensor's acoustic wave signal exhibits a sharp mutation or a specific frequency signal characteristic that occurs simultaneously with a stress-strain mutation, it is considered a wire breakage event. The system automatically calibrates the wire breakage moment and spatial location, establishing a mapping with the corresponding wire number. S3.

22. Determine the quantitative index of damage impact, construct the output characteristic variable, and calculate the relative ratio of the strain amplitude increased by the prestressed steel wire due to wire breakage, which is defined as the stress transfer factor λ s , the calculation formula is as follows: Where, ε after and ε before Respectively represent the strain values ​​of the fiber Bragg grating strain sensor before and after the wire is broken; The maximum radius range of the strain disturbance caused by the broken wire is calculated, that is, the maximum distance from the position of the broken wire where the strain change rate exceeds the threshold η, which is defined as the damage diffusion radius R d , the calculation formula is as follows: Where, ε(r,t after ) is the strain value after the wire is broken at a distance r from the broken wire position; ε(r,t before ) is the strain value before the wire is broken at a distance r from the broken wire position; η is the strain change rate threshold, usually 10% to 20%; δ is a small constant to avoid division by zero, usually 10 -6 ; r is the Euclidean distance from the broken wire position; The ratio of the remaining load-bearing capacity of the structure after wire breakage to the intact state is calculated and defined as the residual load-bearing capacity ratio R c , the calculation formula is as follows: Where p fail is the ultimate bearing pressure value after wire breakage, obtained through pressure test or by finite element simulation or degradation model reasoning; design is the design pressure value; S3.3 Deep Learning Model Training and Prediction S3.

31. Dataset division and standardization. Assuming that a wire break occurs at time t, the input samples ε1, ε2, D3, ε4, and DAS constructed using monitoring data will be DATA and output samples λ s1 ,λ s2 ,λ s3 ,λ s4 、R d5 、R d6 、R d7 、R d8 、R c The test data are standardized. ε1, ε2, ε3, and ε4 represent the strain values ​​of the fiber Bragg grating strain sensors wrapped around the surfaces of the core concrete, steel cylinder, prestressed steel wire, and mortar from tp to t+n, respectively. n and p are integers that can be determined according to the test time and data acquisition frequency. DAS DATA Indicates the value of the DAS sensor before and after the wire is broken, used for wire breakage positioning, λ s1 ,λ s2 ,λ s3 ,λ s4 R represents the stress transfer factor of the core concrete, steel cylinder, prestressed steel wire and mortar. d5 、R d6 、R d7 、R d8 Represents the damage diffusion radius of the core concrete, steel cylinder, prestressed steel wire and mortar; divided into training set, validation set and test set in a ratio of 70%:15%:15%; S3.

32. Determine the multi-objective loss function: Where α is the loss weighting factor, i = 1, 2, 3, 4, j = 5, 6, 7, 8, MSE is the average of the squares of the differences between the predicted value and the actual value, the predicted value is the output value of the multimodal fusion network CNN-GRU, and the actual value is the value measured by the sensor; S3.

33. Use the multimodal fusion network CNN-GRU for modeling. The model includes an input layer, a convolutional layer, a flattening layer, a GRU layer, and an output layer. First, the input data enters the CNN. The CNN extracts local features of the data through convolution and pooling operations to generate a feature map. The feature map output by the CNN is flattened into a one-dimensional vector. The flattened vector is input to the GRU. The GRU processes the sequence data through a gating mechanism and learns time dependencies. Finally, through the fully connected layer and activation function, when the loss function value is less than the threshold or the number of cycles reaches the set value, the prediction result λ is output. s1 ,λ s2 ,λ s3 ,λ s4 、R d5 、R d6 、R d7 、R d8 、R c ; S3.

34. The trained network can be used to intelligently and quantitatively evaluate the impact of PCCP pipe breakage on surrounding structures. That is, by inputting new wire breakage monitoring data into the model, the results of the impact on structural performance can be output in real time.

4. The method for intelligently sensing broken wires in PCCP pipes and testing their impact on structural performance according to claim 3 is characterized in that: The S3.32 is obtained by: (1) When the number of cycles is less than 100, the loss function is weighted and integrated using a preset constant weighting factor, with values ​​of α1-α4 being 1.0, α5-α8 being 0.5, and α9 being 1.5; (2) When the number of cycles is not less than 100, a dynamic loss weighting mechanism is introduced during the training process. Based on the loss value change rate, the relative growth rate of each loss item in the mth round of training is calculated. p is a natural number from 1 to 9, and the weighting factor is dynamically determined based on it: Where, is the loss value of the p-th task during training m-1 iterations.

5. A testing system for intelligent sensing of PCCP pipe corrosion and broken wires and its impact on structural performance according to any one of claims 1 to 4, characterized in that: include: Fiber Bragg grating strain sensors and fiber Bragg grating stress sensors are wound around the surfaces of the core concrete, steel cylinder, prestressed steel wire, and mortar of the PCCP pipe, respectively. DAS sensors are connected in series to the mortar surface and the inner surface of the core concrete. A copper bar is inserted into the mortar as a built-in cathode and led out with a black wire. One end of the red wire is connected to the prestressed steel wire and the other end is connected to the power supply anode. One end of the black wire is connected to the copper bar and the other end is connected to the power supply cathode. The fiber Bragg grating strain sensor, fiber Bragg grating stress sensor, and DAS sensor are connected to the control room signal respectively.

Images