Pccp prestress failure signal acquisition and analysis system and method

Abstract

The present application relates to the field of prestressed concrete cylinder pipe detection, and particularly relates to a PCCP prestress failure signal acquisition and analysis system and method, comprising the following steps: an excitation coil and a magnetic sensor are arranged outside a prestressed steel wire layer of a PCCP pipe, the excitation coil is used to apply a low-frequency alternating excitation magnetic field to the prestressed steel wire layer, the magnetic sensor is used to acquire an eddy current magnetic field signal generated in the prestressed steel wire, and an initial electromagnetic signal of each measuring point is obtained; a strain gauge is fixedly connected between the excitation coil and the magnetic sensor, a resistance signal of the strain gauge is acquired in real time, and a state characteristic value reflecting a mechanical coupling state of a monitoring layer is extracted from the resistance signal. The present application combines the excitation coil, the magnetic sensor and an acousto-optic fiber sensor, cooperatively captures eddy current mutation and stress waves, introduces a strain gauge verification signal to improve reliability, and through confidence management and space-time correlation analysis, cross verification denoising, false negatives and false positives are reduced, and the positioning accuracy and diagnosis depth of broken wires are improved.

Description

Technical Field

[0001] This invention relates to the field of prestressed concrete cylinder pipe testing, and particularly to a PCCP prestressed failure signal acquisition and analysis system and method. Background Technology

[0002] PCCP (Prestressed Concrete Cylinder Pipe) is a composite pipe that combines the prestressing of steel with the pressure-bearing capacity of concrete. It consists of a concrete core wrapped with high-strength prestressed steel wires, a steel cylinder, and an anti-corrosion coating. It boasts advantages such as high crack resistance, high pressure resistance, and long service life, and is widely used in major projects such as long-distance water transmission, urban water supply and drainage, and water conservancy projects. However, during long-term operation, PCCP is susceptible to soil corrosion, external impacts, and pipeline settlement, leading to damage and corrosion of the prestressed steel wires, and ultimately, wire breakage. This reduces the pipeline's load-bearing capacity. If not detected in time, this can cause pipe bursts, resulting in water waste and safety risks. Therefore, developing efficient and accurate PCCP wire breakage detection technology is crucial.

[0003] Currently, monitoring technologies for prestressed steel wire breakage in PCCP mainly include acoustic-fiber optic monitoring and electromagnetic induction monitoring. Electromagnetic induction monitoring relies on the magnetic field of a coil to induce eddy currents in the prestressed steel wire, capturing sudden changes in these eddy currents to pinpoint the location of the broken wire. For example, the method disclosed in Chinese patent application CN111896611A involves pre-embedding an excitation coil and applying low-frequency alternating current to generate an alternating magnetic field. When this magnetic field acts on the prestressed steel wire within the pipe, eddy currents are induced. When the prestressed steel wire breaks, the local electromagnetic structure changes, causing a sudden change in the eddy currents. This signal is captured by a magnetic sensor array, and the breakage location can be accurately pinpointed using a signal inversion algorithm. Acoustic-fiber optic monitoring deploys optical fibers to sense changes in scattered light caused by the stress wave from the broken wire, allowing for precise location of the breakage. For example, the method disclosed in Chinese patent application CN114778678A involves deploying acoustic sensing optical fibers along the pipe axis. The stress wave released at the moment of the prestressed steel wire breakage is transmitted to the optical fiber, causing a slight deformation. By analyzing the phase or frequency changes of the backscattered light from the optical fiber, the breakage event can be identified and located.

[0004] However, both of the aforementioned mainstream detection technologies have significant limitations, making it difficult to meet the monitoring requirements of high precision and high stability. For fiber optic monitoring technology, signal acquisition is highly susceptible to interference from environmental factors such as water flow noise and water hammer impact within the pipeline. These interference signals are similar in characteristics to the stress wave signal from a broken wire, making them difficult to distinguish effectively and resulting in a high false alarm rate. Electromagnetic induction monitoring technology, on the other hand, is limited by the resistance attenuation caused by the length of the prestressed steel wire. As the monitoring distance increases, the magnetic field strength and eddy current signal gradually weaken. In weak signal areas, the signal abrupt changes caused by wire breakage are easily masked by background noise, leading to inaccurate identification of wire breakage in that area and a high risk of missed detection. Summary of the Invention

[0005] This invention provides a PCCP prestress failure signal acquisition and analysis system and method to solve the problem of low accuracy in PCCP wire breakage detection by fiber optic monitoring or electromagnetic induction monitoring.

[0006] To solve the above-mentioned technical problems, this application provides the following technical solution: The method for acquiring and analyzing PCCP prestressed failure signals includes the following steps: An excitation coil and a magnetic sensor are set on the outside of the prestressed steel wire layer of the PCCP pipe. The excitation coil is used to apply a low-frequency alternating excitation magnetic field to the prestressed steel wire layer, and the magnetic sensor is used to collect the eddy current magnetic field signal induced in the prestressed steel wire to obtain the initial electromagnetic signal of each measuring point. A strain gauge is fixedly connected between the excitation coil and the magnetic sensor. The resistance signal of the strain gauge is acquired in real time, and state characteristic values ​​reflecting the mechanical coupling state of the monitoring layer are extracted from the resistance signal. The state characteristic values ​​include at least the baseline drift rate, vibration energy in a specific frequency band, and the count of sudden jump events. An acoustic fiber optic sensor is laid inside the PCCP pipe along the pipe axis to collect the original vibration signal of the entire pipe length in real time. The amplitude attenuation, phase shift, and frequency response variation characteristics of the initial electromagnetic signal are analyzed and extracted. The first wire breakage range and the corresponding electromagnetic confidence level are calculated. The time-frequency characteristics of the original vibration signal and the correlation between the original vibration signal and the state characteristic value are analyzed. The second wire breakage range and the corresponding acoustic confidence level are calculated. When both electromagnetic confidence and acoustic confidence are higher than the preset confidence threshold, and the first and second wire breakage ranges overlap spatially, the overlapping area is determined as a reliable wire breakage location, and the reliable wire breakage location is used as an alarm output; when there is a numerical difference between electromagnetic confidence and acoustic confidence or the spatial ranges do not overlap, the electromagnetic confidence or acoustic confidence is marked in the first and second wire breakage ranges, and the first and second wire breakage ranges are merged as a contradictory alarm output; When only electromagnetic confidence or acoustic confidence is obtained, and the duration of the electromagnetic confidence or acoustic confidence being higher than the preset confidence threshold exceeds the preset judgment threshold, the state feature value is combined to make a judgment and generate a self-diagnostic prompt.

[0007] The basic principle and beneficial effects of this invention are as follows: This invention applies a magnetic field to a prestressed steel wire using an excitation coil, inducing eddy currents within the wire. A magnetic sensor captures the abrupt change in the eddy current magnetic field caused by wire breakage, thus reflecting the electromagnetic continuity of the wire. Simultaneously, an acoustic fiber optic sensor captures the stress wave released at the moment of wire breakage and propagating along the pipe, directly sensing the breakage event. In this process, the monitoring of stress waves and eddy currents is independent and affected by environmental interference mechanisms differently. Therefore, their coordinated response to the same breakage event forms a strong basis for cross-validation, significantly reducing false alarms or missed alarms caused by the limitations of a single technology (only stress wave or eddy current monitoring), thereby improving the accuracy of location. Furthermore, this invention introduces a strain gauge between the excitation coil and the magnetic sensor, sensing the stability of the mechanical coupling state of the monitoring layer itself. The state characteristic values ​​output by the strain gauge provide a crucial self-checking dimension for judging the reliability of electromagnetic and acoustic signals. For example, when the electromagnetic signal displays a high-confidence alarm but the acoustic signal does not detect a broken wire, if the strain gauge characteristics simultaneously indicate a drastic jump in local coupling, it can be inferred that the electromagnetic anomaly likely originates from mechanical interference within the monitoring layer itself rather than a genuine broken wire. This effectively verifies and denoises the data. By accurately acquiring electromagnetic or acoustic-optical data, further verification and diagnosis of the data generation conditions result in higher-quality data ultimately used for fusion decision-making, further ensuring the accuracy of the broken wire location results.

[0008] The baseline drift and other characteristics monitored by strain gauges over a long period of time can reflect the aging or degradation trend of the bond between the excitation coil and the magnetic sensor and the pipeline structure. This allows the present invention to also provide early warning of the failure risk of the monitoring network itself, providing a basis for preventive maintenance.

[0009] Furthermore, the multi-level decision-making rules based on confidence management endow this invention with excellent adaptability and robustness in the face of complex working conditions. This invention outputs hierarchical and categorized conclusions without forcibly synthesizing contradictory information. For example, in situations where strong water flow vibrations cause a temporary decrease in the acoustic signal-to-noise ratio, the confidence weight of the acoustic mode can be automatically reduced, while continuous monitoring is performed using the electromagnetic mode. The source of uncertainty in the current decision is clearly marked in the log, thereby avoiding arbitrary conclusions when information is incomplete and ensuring the prudence of engineering judgments.

[0010] Furthermore, the spatiotemporal correlation analysis of electromagnetic and acoustic signals, combined with the local mechanical state provided by strain gauges, can help distinguish between actual fractures originating from the pipe structure and disturbances such as external impacts. In the case of an actual fracture, the abrupt changes in electromagnetic signals, the generation of acoustic stress waves, and the local stable state indicated by the strain gauges should exhibit a high degree of spatiotemporal consistency. However, in the case of external mechanical impacts, strong correlations may occur in the acoustic and strain gauge signals, but the electromagnetic signals will not show any specific changes. This subtle discrimination capability based on multi-source information correlation patterns enables a more refined identification of the nature of the event, achieving a diagnostic depth that is difficult to obtain by simply improving the accuracy of a single sensor.

[0011] In summary, this invention combines an excitation coil, a magnetic sensor, and an optical fiber acousto-optic sensor to collaboratively capture eddy current mutations and stress waves. It introduces strain gauges to verify signal reliability, and through confidence management and spatiotemporal correlation analysis, cross-validation and noise reduction reduce false alarms and improve the accuracy and depth of wire breakage location.

[0012] Furthermore, the excitation coil is mechanically connected to multiple magnetic sensors through strain gauges, and adjacent magnetic sensors are mechanically connected through strain gauges, thereby forming a sensor network with magnetic sensors and excitation coils as nodes and strain gauges as connecting edges.

[0013] By constructing a sensor network consisting of excitation coils, magnetic sensors, and strain gauges, and winding the network around the outside of the prestressed steel wire layer, forming a bidirectional wrapping structure with the acousto-optic fiber inside the pipe, the steel wire is sandwiched in the middle, enabling omnidirectional deformation capture. The network topology can cross-verify the location of electromagnetic anomalies through strain correlation between adjacent nodes, reducing false alarms due to electromagnetic noise. The bidirectional wrapping structure allows the acousto-optic fiber to capture inward deformation of the steel wire, while even minute outward deformations, such as thermal expansion and contraction, can be accurately captured by the strain gauges, converting physical deformation into electrical signals. Simultaneously, it can identify the tension and contraction changes of the steel wire caused by vibration, adapting to vibration scenarios under different environmental temperatures, such as the minute deformations of pipes contracting at low temperatures in winter and expanding at high temperatures in summer. When water pipelines traverse mountainous areas with large temperature differences, it can avoid misjudging thermal deformation as wire breakage, balancing monitoring sensitivity and environmental adaptability.

[0014] Furthermore, when extracting state feature values ​​from the resistance signal, the resistance signal of each strain gauge is first filtered to separate the dynamic change component and the slowly changing component; the dynamic components of all strain gauges in the sensor network are analyzed by spectrum analysis to obtain the vibration energy of a specific frequency band that reflects the overall state; the rate of change of the slowly changing component of each strain gauge is calculated, and the distribution characteristics of the slowly changing component in the sensor network are statistically analyzed to obtain the baseline drift rate that reflects the overall slowly deforming trend of the monitoring layer. A first time window is set. When the number of strain gauges whose dynamic change component amplitude exceeds a preset first amplitude threshold within the first time window exceeds a preset first number threshold, the time corresponding to the first time window and the position of the strain gauge in the sensor network where the dynamic change component amplitude exceeds the preset first amplitude threshold are recorded as a coordinated mutation event. The frequency of occurrence of coordinated mutation events is counted, and the spatial distribution correlation between strain gauges in the coordinated mutation events is analyzed to obtain a subset of the sudden change event count. A second time window is set. When the dynamic change component amplitude of a single strain gauge exceeds a preset second amplitude threshold within the second time window, the time corresponding to the second time window and the position of the strain gauge in the sensor network are recorded as a sudden change event. The second amplitude threshold is greater than the first amplitude threshold. The number of occurrences of sudden change events within a preset range is counted to obtain another subset of the sudden change event count.

[0015] By separating dynamic and slow components and introducing a dual-time-window detection mechanism, the state characteristic values ​​can simultaneously reflect long-term aging trends and instantaneous mechanical disturbances. The rate of change of the slow component can identify latent faults such as strain gauge aging and adhesion degradation in advance, avoiding missed detections due to sensor failure; the spectral analysis of the dynamic component can capture low-frequency vibration modes in pipeline operation, which can be used to identify background vibrations such as pump resonance and water hammer effects in large-scale water conveyance projects; the dual-time-window mechanism can distinguish between local sudden disturbances and global coordinated vibrations. For example, in external construction, strong impacts at a single node can be identified by the second time window, while synchronous vibrations at multiple nodes can be captured by the first time window, thereby improving the accuracy of event classification.

[0016] Furthermore, when analyzing and extracting the characteristics of the initial electromagnetic signal and calculating the electromagnetic confidence score, the initial electromagnetic signals collected by all magnetic sensors in the sensor network are first acquired. The initial electromagnetic signal of each magnetic sensor is compared with the corresponding reference data, and the attenuation ratio, phase shift, and frequency response difference are calculated. Initial electromagnetic signals with differences exceeding a preset normal threshold are considered abnormal signals. Then, the spatial distribution and aggregation pattern of the magnetic sensors that acquired the abnormal signals on the sensor network topology are analyzed. The first broken wire range and electromagnetic confidence score are calculated based on the spatial distribution and aggregation pattern. The reference data is the baseline data of the electromagnetic signal collected by the corresponding magnetic sensor when the prestressed steel wire of the PCCP pipe is in a complete and undamaged state.

[0017] By comparing electromagnetic signals with reference data and performing anomaly clustering analysis, wire breakage detection achieves high resolution and robustness. Multi-dimensional comparisons of amplitude attenuation, phase shift, and frequency response differences effectively distinguish between wire breakage and sensor aging. Spatial clustering analysis of abnormal signals avoids misjudging occasional noise from a single magnetic sensor as wire breakage. When pipelines traverse industrial parks (such as chemical plants or machining areas), transient electromagnetic pulses generated by the start-up and shutdown of surrounding equipment can easily interfere with sensor signals. This method uses clustering patterns to determine the spatial consistency of the interference, identifying only electromagnetic anomalies occurring synchronously at multiple nodes, eliminating single-point transient interference, and ensuring stable and reliable wire breakage detection in complex industrial electromagnetic environments.

[0018] Furthermore, when analyzing the time-frequency characteristics of the original vibration signal and the correlation between the original vibration signal and the state characteristic value, the original vibration signal is first transformed by time and frequency. The sequence of signal energy changing with time within the preset fracture frequency range is extracted as the characteristic frequency band energy time history. The spatiotemporal distribution of the characteristic frequency band energy time history is matched with the spatial distribution of vibration energy in a specific frequency band obtained from the sensor network. The matching degree coefficient is calculated. Combining the intensity, spatiotemporal focus, and matching degree coefficient of the characteristic frequency band energy time history, the range of the second broken wire and the acoustic confidence level are calculated.

[0019] The characteristic frequency band energy time history can accurately capture the propagation of stress waves at the moment of wire breakage, and matching with vibration energy in a specific frequency band can further verify the consistency of the signal source. In long-distance PCCP pipelines, this invention can effectively distinguish the stress wave of wire breakage from background signals such as water flow noise and mechanical vibration. By introducing the matching coefficient, the acoustic detection results can be mutually verified with the mechanical state, significantly improving the reliability of wire breakage judgment.

[0020] Furthermore, when there are numerical differences or spatial non-overlapping ranges between electromagnetic confidence and acoustic confidence, the dynamic change components of strain gauges corresponding to adjacent magnetic sensors in the sensor network are first obtained; cross-correlation calculations are performed on the dynamic change components of adjacent strain gauges to obtain correlation coefficients used to characterize the similarity and synchronous change trend of the two dynamic change components; the maximum absolute value or root mean square value of the dynamic change component within the first or second time window is used as the amplitude of the dynamic change component, and the correlation coefficient and the amplitude of the dynamic change component of the strain gauge are used for time synchronization verification to obtain the disturbance synchronization index of adjacent nodes; based on the correlation coefficient and synchronization index, a spatial consistency map characterizing the distribution of external vibration interference is constructed; for the regions in the first and second broken wire ranges where the synchronization index exceeds a preset synchronization threshold, the weights of electromagnetic confidence and acoustic confidence are reduced to obtain the corrected first and second broken wire ranges; then, electromagnetic confidence or acoustic confidence is marked in the first and second broken wire ranges.

[0021] By analyzing the cross-correlation of dynamic components and verifying synchronicity, a spatial consistency map is constructed, effectively solving the problem of inconsistency between electromagnetic and acoustic confidence levels. Cross-correlation calculations can identify synchronous vibrations of adjacent strain gauges, thereby determining whether external disturbances have spatial consistency. The synchronicity index can be used to correct the range of wire breakage, avoiding misjudging strong vibration areas as wire breakage. When piling work is carried out above pipelines, this invention can identify global vibration disturbances through the consistency map, thus protecting the wire breakage detection results from the influence of construction noise.

[0022] Furthermore, when only electromagnetic or acoustic confidence levels are obtained, and the duration for which the electromagnetic or acoustic confidence levels exceed a preset confidence threshold exceeds a preset judgment threshold, the abnormal signals of all magnetic sensors within the corresponding first or second broken wire range are first extracted. The attenuation ratio, phase shift, and frequency response difference corresponding to the abnormal signals are obtained to obtain a feature difference matrix used to characterize the spatial distribution differences of the abnormal signals. The amplitude of the dynamic change components of adjacent strain gauges is obtained, and correlation analysis is performed in conjunction with the feature difference matrix to obtain the abnormal disturbance correlation coefficient. Based on the feature difference matrix, it is determined whether the abnormal signal is concentrated in a single magnetic sensor node or is continuously distributed along the sensor network topology. The judgment result is processed into an abnormal location, and then combined with the abnormal disturbance correlation coefficient, a self-diagnostic prompt pointing to the abnormal location is generated.

[0023] Automatic diagnosis is achieved through correlation analysis between the characteristic difference matrix of abnormal signals and the dynamic components of strain. The characteristic difference matrix clearly presents the spatial distribution of abnormal signals, thereby determining whether the abnormality is a local sensor failure or structural damage; the correlation with the amplitude of the dynamic components can further distinguish between mechanical disturbances and sensor failure. When a magnetic sensor experiences baseline drift due to moisture, this invention can identify its isolation through the difference matrix and confirm the absence of mechanical disturbance by combining strain data, thereby automatically generating a sensor fault indication and avoiding the tedious manual troubleshooting.

[0024] Furthermore, the resistance signals of all strain gauges in the sensor network are first acquired, and the resistance signals are processed using a spatial interpolation algorithm to obtain the overall deformation distribution field. Gradient calculation is performed on the overall deformation distribution field to obtain the deformation gradient differences at different locations along the vertical height of the PCCP pipe. Based on these deformation gradient differences, the longitudinal uniformity of the pipe stiffness distribution is assessed. The dynamic variation components of all strain gauges in the sensor network are extracted, and spectral analysis is performed on these components to obtain the power spectral density of each strain gauge's dynamic components. From the power spectral density of each strain gauge, the frequency component with the largest power value is selected as the local dominant frequency of a single strain gauge. The local dominant frequencies of all strain gauges are statistically averaged to obtain the dominant frequency corresponding to the vibration energy of a specific frequency band in the overall sensor network. Dominant frequency data at different time points are continuously collected, and the change in dominant frequency over time is calculated to obtain the vibration dominant frequency migration trend, which characterizes the evolution trend of the vibration spectrum. The vibration dominant frequency migration trend is correlated and verified with the trend of uniformity of pipe stiffness distribution to obtain a trend coordination coefficient. When the trend coordination coefficient exceeds a preset coordination threshold, a structural performance degradation early warning index is generated.

[0025] By analyzing the correlation between the overall deformation distribution field and the migration of the dominant vibration frequency, a long-term assessment of the overall health status of the pipeline can be achieved. The overall deformation distribution field can intuitively present the changes in the longitudinal (vertical) stiffness of the pipeline, providing early warnings of pipeline aging or local structural deterioration. The migration of the dominant vibration frequency reflects the decrease in the overall stiffness of the pipeline, and its correlation with the deformation field can verify whether the deterioration is structural. In large-scale water supply projects, when the concrete layer of the pipeline experiences fatigue due to long-term water pressure changes, this invention can identify early degradation through the coordinated changes in the dominant frequency migration and deformation gradient, thus avoiding sudden wire breakage accidents.

[0026] Furthermore, if the deformation corresponding to each spatial coordinate in the overall deformation distribution field exceeds the normal deformation threshold, the deformation gradient difference exceeds the preset normal gradient threshold, or the difference between the overall deformation distribution field and the preset normal distribution model exceeds the preset difference threshold, then the overall deformation distribution field is decomposed into a direction vector to obtain a deformation direction vector characterizing the deformation direction; the deformation direction vector is statistically analyzed to calculate the main force direction of the pipe; the overall deformation distribution field is spatially attenuated to obtain a deformation attenuation coefficient, and the stress transmission path is determined based on the deformation attenuation coefficient; the characteristic frequency band energy time history of the optical fiber sensor is extracted, and the propagation speed of the characteristic frequency band energy time history is fitted to obtain the vibration event propagation direction; the angle between the main force direction and the vibration event propagation direction is calculated to obtain a direction consistency coefficient characterizing the consistency of the two directions; when the direction consistency coefficient exceeds the preset direction threshold and the duration exceeds the preset duration threshold, the anomaly is determined to originate from a directional external disturbance; combining the abnormal area of ​​the overall deformation distribution field and the propagation range of the vibration event, early warning information including the disturbance location and the affected area is generated.

[0027] By analyzing the consistency between the deformation direction vector and the vibration propagation direction, accurate identification of directional external disturbances can be achieved. The deformation direction vector reveals the dominant direction of force on the pipeline, while comparison with the vibration propagation direction can determine whether the disturbance originates from the outside. When excavation or piling is being carried out near the pipeline, this invention can accurately identify the direction of disturbance origin and assess its impact range. Through continuous monitoring of the direction consistency coefficient, it can effectively avoid misjudging external construction disturbances as wire breakage, significantly improving the reliability of this invention in complex environments. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the PCCP pipe structure; Figure 2 This is a flowchart of the PCCP prestress failure signal acquisition and analysis method in Example 1; Figure 3 This is a schematic diagram of the structure between the prestressed steel wire, the excitation coil, and the magnetic sensor. Figure 4 This is a schematic diagram of the warning information including the PCCP pipe number and the location of the broken wire in Example 1; The reference numerals in the attached figures include: 1. Sand and gravel mixture outer layer; 2. Sensor layer; 3. Prestressed steel wire; 4. Structural steel cylinder; 5. Concrete inner layer; 6. Excitation coil; 7. Magnetic sensor; 8. Strain gauge. Detailed Implementation

[0029] The following will describe the concept and technical effects of the present invention clearly and completely with reference to embodiments, so as to fully understand the purpose, features and effects of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are all within the scope of protection of the present invention. Example 1 like Figure 1 As shown, the PCCP pipe consists of, from the inside out, an inner concrete layer 5, a structural steel cylinder 4, a prestressed steel wire layer, a sensor layer 2, and a sand and gravel mixed outer layer 1.

[0030] This embodiment includes a PCCP prestressing failure signal acquisition and analysis method, and a PCCP prestressing failure signal acquisition and analysis system using the PCCP prestressing failure signal acquisition and analysis method. The PCCP prestressing failure signal acquisition and analysis system includes an acquisition module and a processor. The processor acquires various data from the acquisition module and executes the PCCP prestressing failure signal acquisition and analysis method.

[0031] like Figure 2 As shown, the PCCP prestress failure signal acquisition and analysis method includes the following steps: An excitation coil 6 and a magnetic sensor 7 are installed on the outside of the prestressed steel wire layer of the PCCP pipe. The magnetic sensor 7 is a high-sensitivity Hall sensor. The excitation coil 6 is made of multi-strand copper core enameled wire, typically with 50-100 turns. The diameter of the excitation coil 6 is adapted to the detection range of the magnetic sensor 7. The excitation coil 6 is electrically connected to an excitation circuit, which is electrically connected to a main control computer (a large industrial-grade main control computer with multi-channel synchronous acquisition, real-time signal processing, and network topology calculation capabilities, which can meet the online monitoring needs of long-distance PCCP pipelines).

[0032] The main control computer stores the positions of each excitation coil 6 within the PCCP pipe (with the pipe axis as the X-axis, the circumferential direction as the Y-axis, and the radial direction as the Z-axis). In this embodiment, this position is used as the excitation position. The main control computer controls the frequency and magnitude of the AC current input to the excitation coil 6 corresponding to the excitation position through the excitation circuit. The AC current is in the form of... .in The excitation current amplitude is set by the main control computer based on the pipe diameter and the thickness of the steel wire layer. Generally, when the PCCP pipe diameter is 1.2-1.5m, it is set to 0.8-1.2A. For the excitation frequency, this embodiment uses low-frequency excitation, which is generally in the range of 1 to 10 kHz. In this embodiment, 5 kHz is selected by default to ensure that the eddy current penetration depth covers the prestressed steel wire layer. This is the initial phase, with a default value of 0-2. The excitation circuit is powered by an internal crystal oscillator. For time synchronization, the main control computer uses a high-precision clock. for The magnitude of the alternating current at that moment. Excitation coil 6 is used to apply a low-frequency alternating excitation magnetic field to the prestressed steel wire layer. In this embodiment, it is assumed that the magnetic field strength range is 5-20mT to ensure that a stable eddy current is induced in the steel wire layer.

[0033] Magnetic sensor 7 is used to collect the eddy current magnetic field signal induced in the prestressed steel wire 3. The main control computer is electrically connected to each magnetic sensor 7 through the magnetic sensor 7 acquisition module. The magnetic sensor 7 acquisition module stores the position of each magnetic sensor 7 in the PCCP tube. The magnetic sensor 7 acquisition module correlates the initial electromagnetic signal (voltage signal) measured by each magnetic sensor 7 with the acquisition time and the position of the magnetic sensor 7, and sends it to the main control computer after analog-to-digital conversion. The main control computer stores a map containing the setting positions of each PCCP tube segment (e.g., ...). Figure 4 As shown in the image, the map contains the numbers of each PCCP pipe section, such as... Figure 4 The corresponding numbers are "56" to "61". The main control computer associates the position of magnetic sensor 7 with the excitation position.

[0034] A strain gauge 8 is fixedly connected between the excitation coil 6 and the magnetic sensor 7, such as... Figure 3 As shown, the prestressed steel wire 3 is embedded within the prestressed steel wire layer. The excitation coil 6 and the magnetic sensor 7 are embedded within the sensor layer 2. The strain gauge 8 is positioned between the excitation coil 6 and the magnetic sensor 7, and is fixedly bonded to them by high-temperature resistant epoxy resin adhesive. It is also insulated from the excitation coil 6 and the magnetic sensor 7 by a polyimide film to prevent electromagnetic interference with the resistance signal acquisition. The strain gauge 8 is made of a long strip of metal, with a sensitivity coefficient of 2.0±0.5% and an operating temperature range of -40℃ to 85℃, making it suitable for outdoor pipeline monitoring environments.

[0035] The excitation coil 6 is mechanically connected to multiple magnetic sensors 7 via strain gauges 8. Adjacent magnetic sensors 7 are mechanically connected to each other via strain gauges 8, thus forming a sensor network with magnetic sensors 7 and excitation coil 6 as nodes and strain gauges 8 as connecting edges. The number of magnetic sensors 7 and excitation coil 6 can be increased according to the network node density set by the administrator based on the monitoring accuracy. The main control computer acquires the resistance signals of each strain gauge 8 through the strain acquisition module (synchronized with the magnetic sensor 7 acquisition module). The resistance signals and electromagnetic signals are acquired synchronously. During the acquisition process, a temperature compensation algorithm (compensation range -40℃ to 85℃) is used to eliminate the influence of ambient temperature on the resistance signals. The sensor network is wound around the outside of the prestressed steel wire layer to capture the outward deformation of the steel wire (such as thermal expansion and contraction, radial expansion caused by external extrusion); the fiber optic sensor inside the tube is laid axially in the inner concrete layer 5 to capture the inward deformation of the steel wire (such as shrinkage under stress, radial shrinkage caused by extrusion of the concrete layer). The two work together to achieve omnidirectional and minute deformation monitoring of the steel wire layer.

[0036] The resistance signal of strain gauge 8 is acquired in real time, and state feature values ​​reflecting the mechanical coupling state of the monitoring layer are extracted from the resistance signal. The state feature values ​​include at least the baseline drift rate, vibration energy in a specific frequency band, and the count of sudden jump events. When extracting state feature values ​​from the resistance signal, the resistance signal of each strain gauge 8 is first filtered to separate the dynamic change component and the slowly changing component.

[0037] Specifically, a second-order Butterworth filter is used to separate the resistance signal. A low-pass filter is set with a cutoff frequency corresponding to the low-frequency band to extract slowly changing components, primarily corresponding to slow-changing factors such as temperature drift, sensor aging, and long-term deformation. A high-pass filter is set with a cutoff frequency corresponding to the mid-to-high-frequency band to extract dynamically changing components, primarily corresponding to fast-changing factors such as vibration and instantaneous deformation. The filtering process is implemented through recursive calculation. The output signal is obtained by weighted summation of the current and two previous input signals, and the output signals from the two previous time steps. The low-pass and high-pass filters have different input signal weighting coefficients, while the recursive coefficients of the output signal are uniformly set. The sampling period is used to define the intervals of the signal in the time series, ensuring the temporal consistency of the filtering calculation.

[0038] Spectral analysis is performed on the dynamic components of all strain gauges 8 within the integrated sensor network to obtain the vibration energy in a specific frequency band reflecting the overall state. Specifically, the spectral analysis is performed on the dynamic variation components of all strain gauges 8 within the integrated sensor network. A Fast Fourier Transform (FFT) is used to convert the time-domain signal to a frequency-domain signal. The number of transform points is set according to actual needs to ensure that the frequency resolution meets the requirements of subsequent analysis. The frequency-domain signal is obtained by sequentially substituting the sampled values ​​of the dynamic components in the time series into the transform process. The transform result reflects the distribution of the signal at different frequencies. Based on the typical characteristics of pipe vibration and wire tension / tension, a frequency range containing the main vibration components is pre-selected as a specific frequency band. The vibration energy in the specific frequency band is obtained by calculating the sum of the squares of the amplitudes corresponding to all frequency points within that frequency band. This energy value is used to characterize the overall vibration intensity of the sensor network within that frequency band.

[0039] The rate of change of the slow-changing components of each strain gauge 8 is calculated, and the distribution characteristics of these slow-changing components in the sensor network are statistically analyzed to obtain the baseline drift rate, which reflects the overall slow deformation trend of the monitoring layer. Specifically, the rate of change of the slow-changing components of each strain gauge 8 is calculated by the ratio of the difference between the slow-changing components at two adjacent moments to the time interval. This time interval is reasonably set according to the response characteristics of the slow-changing factors to ensure accurate capture of long-term changing trends. The main control computer uses the Kriging interpolation algorithm to statistically analyze the spatial distribution characteristics of the slow-changing components of all strain gauges 8 in the sensor network. By interpolating, the change state of the network coverage area is completed, thereby obtaining the overall slow deformation trend of the monitoring layer, which is the baseline drift rate. The baseline drift rate is used to characterize latent degradation phenomena such as sensor aging and long-term pipeline deformation. It can detect slow degradation processes that are not easily detected in advance, eliminating hidden interference for subsequent fault diagnosis.

[0040] A first time window is defined. When the number of strain gauges 8 whose dynamic change component amplitude exceeds a preset first amplitude threshold within the first time window exceeds a preset first number threshold (set by the administrator based on the amplitude of minor vibrations in the PCCP tube), the time corresponding to the first time window and the location of the strain gauges 8 in the sensor network whose dynamic change component amplitude exceeds the preset first amplitude threshold (set by the administrator based on the proportion of minor vibration coordinated response, for example, 30% of the total number of sensor nodes in a certain PCCP tube) are recorded as coordinated mutation events. The frequency of occurrence of coordinated mutation events is counted, and the spatial distribution correlation between strain gauges 8 in coordinated mutation events is analyzed to obtain a subset of the sudden jump event count (corresponding to global coordinated vibration events). Among them, spatial clustering algorithms (e.g., Euclidean distance clustering, threshold 1m) are used to analyze the spatial distribution correlation between strain gauges 8.

[0041] A second time window is set. When the amplitude of the dynamic change component of a single strain gauge 8 exceeds a preset second amplitude threshold (greater than the first amplitude threshold, corresponding to a local strong impact) within the second time window, the time corresponding to the second time window and the position of the strain gauge 8 in the sensor network are recorded as a sudden change event. The number of sudden change events occurring within a preset range is counted to obtain another subset of the sudden change event count (corresponding to local sudden disturbance events).

[0042] The administrator sets the duration of the first and second time windows based on the different response characteristics of pipeline vibration and steel wire deformation. The first time window is used to capture weak vibrations occurring simultaneously at multiple nodes. These vibrations are usually caused by changes in the overall stress on the pipeline, external environmental disturbances, or slight tension or contraction of the steel wire layer. Their changes are slow and their duration is long, therefore a relatively long time window is needed to ensure sufficient sampling points are covered and to accurately identify whether multiple strain gauges 8 show synchronous changes within the same time period. The second time window is used to capture sudden, strong impacts or instantaneous deformations. These events are usually short in duration and concentrated in energy, therefore a relatively short time window is needed to avoid signal averaging and ensure timely capture of the instantaneous jumps in individual strain gauges 8.

[0043] An optical fiber acoustic sensor is installed along the axial direction inside the PCCP pipe to collect the original vibration signal of the entire pipe length in real time. A distributed optical fiber acoustic sensor is selected, and the optical fiber is a single-mode fiber.

[0044] The amplitude attenuation, phase shift, and frequency response characteristics of the initial electromagnetic signal are analyzed and extracted. The range of the first broken wire and the electromagnetic confidence level corresponding to the range of the first broken wire are calculated.

[0045] Specifically, the initial electromagnetic signals collected by all magnetic sensors 7 in the sensor network are first acquired, and the current signal of each magnetic sensor 7 is compared with the corresponding reference data. The reference data is obtained by calibrating and collecting the electromagnetic signals when the prestressed steel wire 3 is in an intact and undamaged state before the pipeline is laid and put into use. It is stored in the main control computer and is calibrated and updated periodically according to the actual operation.

[0046] During the comparison process, signal differences were calculated from three aspects. First, the amplitude attenuation was assessed by comparing the amplitude of the current signal with that of the reference signal to determine if the signal had weakened. Second, the phase shift was assessed by extracting and comparing the phase information of the current signal and the reference signal to determine if the signal phase had shifted. Third, the frequency response characteristics were assessed by analyzing the response characteristics of the current signal and the reference signal at different frequencies to determine if their frequency domain characteristics had changed.

[0047] Based on the signal characteristics under normal pipeline operation, a normal variation range is preset. If any of the above three differences exceeds the preset normal range, the initial electromagnetic signal collected by the magnetic sensor 7 is considered an abnormal signal.

[0048] A spatial density clustering algorithm is used to analyze all magnetic sensors 7 that have collected abnormal signals. Based on the positional relationship of each sensor in the sensor network topology, the spatial distribution and clustering characteristics of the abnormal signals are determined. By identifying whether sensors with abnormal signals form continuous clustering areas in space, the potential range of wire breakage is determined. The boundary of this clustering area is the first wire breakage range, and its location is marked according to the three-dimensional coordinate system of the pipeline to ensure accurate positioning. Electromagnetic confidence is used to measure the reliability of the wire breakage determination result. It is obtained by comprehensively considering the differences in amplitude attenuation, phase shift, and frequency response characteristics of the abnormal signals. These three differences are first converted into relative quantities within a unified range, and then comprehensively calculated according to pre-set importance weights. The higher the degree of difference in amplitude attenuation, phase shift, and frequency response characteristics, the closer the corresponding relative quantity is to the end indicating anomaly, and the higher the comprehensive electromagnetic confidence. The electromagnetic confidence value ranges from no anomaly to highly anomaly; the closer it is to the anomaly end, the greater the probability of wire breakage, and the more reliable the determination result.

[0049] Analyze the time-frequency characteristics of the original vibration signal and the correlation between the original vibration signal and the state characteristic value, and calculate the second wire breakage range and the acoustic confidence level corresponding to the second wire breakage range.

[0050] Specifically, the original vibration signal is first transformed by time and frequency, and the sequence of signal energy changing with time that conforms to the preset fracture frequency range is extracted as the characteristic frequency band energy time history. The spatiotemporal distribution of the characteristic frequency band energy time history is matched with the spatial distribution of vibration energy in a specific frequency band obtained from the sensor network, and the matching degree coefficient is calculated. Combining the intensity, spatiotemporal focus and matching degree coefficient of the characteristic frequency band energy time history, the range of the second broken wire and the acoustic confidence level are calculated.

[0051] More specifically, the original vibration signal is first subjected to time-frequency transformation using wavelet transform to convert the time-domain signal into a two-dimensional time-frequency matrix that simultaneously contains time and frequency information, thus enabling observation of the energy distribution of the signal at different times and frequencies. Based on the typical frequency characteristics of the stress wave generated by the broken wire, a frequency range containing the main energy of the broken wire is pre-selected. Within this frequency range, a sequence of signal energy changes over time is extracted; this sequence is the characteristic frequency band energy time history, used to reflect the propagation process of the broken wire stress wave within the pipe.

[0052] The temporal and spatial distribution of characteristic frequency band energy time histories is compared with the spatial distribution of vibration energy in a specific frequency band obtained from a sensor network to analyze the similarity and correspondence between the two. A matching coefficient is obtained by calculating the correlation between the two energy sequences; this coefficient measures the consistency between the vibration signal and the strain monitoring results. A higher matching coefficient indicates that the vibration sources reflected by the two monitoring methods are more likely to be the same.

[0053] Based on this, the acoustic confidence level is calculated by combining the intensity of the characteristic frequency band energy time history, the concentration of energy in time and space, and the matching coefficient. The acoustic confidence level is used to determine the reliability of the fiber optic monitoring results indicating a broken wire; the closer the value is to the end indicating high reliability, the greater the probability of a broken wire. The second broken wire range is determined based on the spatial location of the strongest energy in the characteristic frequency band energy time history, and the range is marked in conjunction with the spatial resolution capability of the optical fiber.

[0054] When both electromagnetic and acoustic confidence levels reach a preset high level of reliability, and the first and second wire breakage areas significantly overlap spatially, the overlapping area is determined to be a reliable wire breakage location. The main control computer will then display the wire breakage location and the corresponding pipe number (e.g., ...). Figure 4 As shown, the location of the broken wire (near pipe number "57") is marked with a red dot, and the warning information along with the relevant confidence level information is sent to the monitoring center via the network, while simultaneously triggering a local audible and visual alarm to prompt staff to handle the situation promptly.

[0055] When there is a numerical difference between the electromagnetic confidence level and the acoustic confidence level (here, one is greater than the administrator's preset value, and the other is lower than the administrator's preset value, for example, one is ≥0.7 and the other is <0.7) or the spatial ranges do not overlap (for example, the overlap area is <30%), the dynamic change components of the strain gauges 8 corresponding to adjacent magnetic sensors 7 in the sensor network are first obtained; cross-correlation calculation is performed on the dynamic change components of adjacent strain gauges 8 to obtain the correlation coefficient used to characterize the similarity and synchronous change trend of the two dynamic change components. A time delay parameter is introduced in the analysis process. By traversing all possible values ​​within a reasonable delay interval, the delay value that maximizes the similarity between the two components is found, and the correlation coefficient is then calculated based on this delay value. The correlation coefficient is used to characterize the synchronous change trend of the two sets of dynamic components. The stronger the similarity and synchronicity, the closer the correlation coefficient is to the end representing the high degree of consistency.

[0056] The maximum absolute value or root mean square value of the dynamic change component within the first or second time window is used as the amplitude of the dynamic change component. The correlation coefficient is then used to verify the time synchronization of the dynamic change component amplitude with strain gauge 8, resulting in an adjacent node disturbance synchronization index. The synchronization verification uses the extreme value of the dynamic change component within the corresponding time window as the component amplitude, combined with the aforementioned correlation coefficient, to verify the time synchronization of the dynamic changes of adjacent strain gauges 8. The verification process must consider the amplitude difference between the two components to mitigate misjudgments of synchronization caused by excessive amplitude deviation, ultimately yielding the adjacent node disturbance synchronization index. The closer this index value is to the end representing high synchronization, the more likely the disturbances reflected by the two components originate from events occurring at the same time.

[0057] Based on the correlation coefficient and synchronicity index, a spatial consistency map is constructed using a heat map drawing method to intuitively present the distribution of external vibration interference in each region of the pipeline. The color change of the map corresponds to the difference in the synchronicity index.

[0058] Before performing confidence level correction, initial weights need to be set for the electromagnetic and acoustic confidence levels. These initial weights are determined based on the inherent reliability of the two monitoring methods. Electromagnetic monitoring is sensitive to changes in the wire structure and is assigned a higher initial weight; acoustic monitoring responds rapidly to instantaneous stress waves but is easily affected by noise and is assigned a lower initial weight. The initial weights can be adjusted appropriately based on the engineering environment, sensor layout, and historical data to ensure the stability and accuracy of wire breakage detection.

[0059] For regions within the first and second wire breakage ranges where the synchronization index exceeds a preset synchronization threshold, external vibration interference is identified, and the weighting of electromagnetic and acoustic confidence in that region is reduced accordingly. For regions where the index is within the normal range, the original confidence weights are maintained, thus obtaining the corrected confidence levels and the corrected first and second wire breakage ranges. Subsequently, the corrected confidence levels are marked in both wire breakage ranges, and the two ranges are integrated using a spatial fusion algorithm to generate and output a contradiction warning. The warning information indicates the possibility of external interference and the average level of the synchronization index, providing a reference for staff to make judgments.

[0060] When only electromagnetic or acoustic confidence scores are obtained (e.g., one has a valid value, the other has no data due to sensor failure), and the duration of the electromagnetic or acoustic confidence scores being higher than the preset confidence threshold (assumed to be 0.7) exceeds the preset judgment threshold (set by the administrator, assumed to be 1 minute), firstly, extract the abnormal signals of all magnetic sensors 7 within the corresponding first or second broken wire range, obtain the attenuation ratio, phase shift, and frequency response difference corresponding to the abnormal signals, and obtain the feature difference matrix used to characterize the spatial distribution difference of the abnormal signals. , As shown below: , This refers to the number of magnetic sensors 7 within this range; For the first The amplitude attenuation ratio of each magnetic sensor 7; For the first Phase shift of magnetic sensor 7; For the first The frequency response differences of the magnetic sensors 7.

[0061] Eigendary difference matrix The study fully presents the spatial distribution and multidimensional characteristics of electromagnetic anomalies within the area of ​​the broken wire, providing basic data for subsequent determination of the source of the anomalies.

[0062] Obtain the dynamic variation component amplitude of adjacent strain gauge 8 Combined with the feature difference matrix Correlation analysis was performed to calculate the correlation coefficient of the abnormal disturbance. , The calculation formula is as follows: , This is the j-th anomalous feature of the i-th magnetic sensor 7; The amplitude of the dynamic component corresponding to the i-th position; The numerator of the calculation formula is the sum of the covariances of the anomalous features and the mechanical disturbances, reflecting the overall correlation between the two; the denominator is the product of the moduli of the two data points, used for normalization. The range remains between -1 and 1. The closer it is to 1, the higher the correlation between electromagnetic anomalies and mechanical disturbances, indicating that the anomaly is very likely caused by structural damage (such as broken wires); The closer to 0, the weaker the correlation between electromagnetic anomalies and mechanical disturbances, which may be due to an anomaly in the magnetic sensor 7 itself; A value closer to -1 indicates a negative correlation, generally caused by noise or measurement error. The correlation coefficient is determined by abnormal disturbances. This can effectively distinguish between structural damage and sensor malfunction, improving the self-diagnostic capability of the present invention.

[0063] The system determines whether the abnormal signal is concentrated at a single magnetic sensor node 7 or is continuously distributed along the sensor network topology based on the feature difference matrix, and processes the results as anomaly locations. Specifically, a clustering analysis algorithm is used to analyze the feature difference matrix, and the source of the anomaly is determined based on the distribution pattern of the abnormal signal in the sensor network. If the analysis results show that the abnormal signal is concentrated at a single magnetic sensor node 7, while the difference indicators of other nodes are within the normal range, it is determined that the sensor itself has a local fault. If the abnormal signal shows a continuous distribution characteristic on the sensor network topology, and the difference indicators of multiple adjacent nodes show gradient changes, it is determined that there is damage to the pipeline structure.

[0064] Then, by combining the correlation coefficient of abnormal disturbances, a self-diagnostic prompt pointing to the location of the anomaly is generated. Specifically, the abnormal areas obtained from cluster analysis are converted into specific location information, which can be labeled according to sensor number or pipeline 3D coordinates. Based on the correlation coefficient of abnormal disturbances, a corresponding self-diagnostic prompt is generated. If the correlation coefficient is low, it indicates that the electromagnetic anomaly and mechanical disturbance are not significantly related, and the prompt is a sensor malfunction; if the correlation coefficient is high, it indicates that the electromagnetic anomaly and mechanical disturbance occur simultaneously, and the prompt is structural damage. Finally, the self-diagnostic prompt, including the location of the anomaly and the judgment criteria, is output to the monitoring center for further processing by staff.

[0065] This embodiment also includes a PCCP prestressing failure signal acquisition and analysis system that uses the PCCP prestressing failure signal acquisition and analysis method.

[0066] In specific implementation, the PCCP pipe consists of, from the inside out, a concrete inner layer 5, a structural steel cylinder 4, a prestressed steel wire layer, a sensor layer 2, and a sand and gravel mixed outer layer 1. This embodiment utilizes multiple types of sensors and signal processing algorithms to achieve prestressing failure monitoring. The specific steps are as follows: An excitation coil 6 and a high-sensitivity Hall sensor are arranged on the outside of the steel wire layer. The coil is wound with multi-strand copper core enameled wire, and the number of turns is adapted to the sensor's detection range. It is connected to a large industrial-grade main control computer via an excitation circuit. The main control computer locates the coil position according to the three-dimensional coordinates of the pipeline, outputs low-frequency alternating current to generate an alternating magnetic field, induces stable eddy currents in the steel wire layer, and simultaneously acquires the initial electromagnetic signal from the magnetic sensor 7 through the acquisition module. After associating the position and time information, the data is stored and analyzed.

[0067] A long strip of metal strain gauge 8 is bonded between the coil and the sensor, insulated by a polyimide film. The strain gauges 8 connect to form a network with nodes representing the coil and sensor, and edges representing the strain gauges 8. The node density can be adjusted as needed. The main control computer synchronously acquires the resistance signal, performs temperature compensation and filtering separation, extracts dynamic and slowly changing components, and then obtains three types of characteristic values—vibration energy in a specific frequency band, baseline drift rate, and sudden jump event count—through spectrum analysis, rate of change calculation, and dual-time-window monitoring.

[0068] A single-mode fiber optic acoustic sensor is axially installed inside the pipe to collect the original vibration signal and perform time-frequency transformation. The energy time history of the characteristic frequency band is extracted and matched with the vibration energy of the strain network to calculate the acoustic confidence level and the range of the second broken wire. At the same time, the range of the first broken wire and the electromagnetic confidence level are obtained by comparing the electromagnetic signal with the benchmark data and performing cluster analysis.

[0069] When both confidence levels meet the standard and their ranges overlap, a reliable fault location is output and an alarm is triggered. If they are inconsistent, an interference map is constructed through cross-correlation calculation and synchronization verification. After correcting the confidence levels according to the initial weights (with higher weights for electromagnetic monitoring), a contradictory warning is output. When only one confidence level meets the standard, a feature difference matrix is ​​constructed, the correlation coefficient of abnormal disturbances is calculated, and cluster analysis is used to distinguish between sensor faults and structural damage. A precise self-diagnostic prompt is generated and sent to the monitoring center.

[0070] Example 2 The only difference between this embodiment and embodiment 1 is that the resistance signals of all strain gauges 8 in the sensor network are first obtained, and the discrete resistance signals are calculated using the Kriging space interpolation algorithm to complete the deformation data of the sensor network coverage area, generate a continuous overall deformation distribution field, and obtain the overall deformation distribution field, which intuitively presents the magnitude and distribution law of deformation at each position of the tube.

[0071] Gradient calculations are performed on the overall deformation distribution field, focusing on the vertical height direction of the PCCP pipeline. The differences in deformation gradients at different radial and circumferential positions are extracted. Based on the differences in deformation gradients, the longitudinal uniformity (longitudinal consistency) of the pipe stiffness distribution is evaluated. The smaller the gradient difference, the more uniform the pipe stiffness distribution and the better the structural integrity; conversely, it indicates local stiffness weakening, which may indicate initial degradation.

[0072] The dynamic variation components of all strain gauges 8 in the sensor network are extracted, and spectral analysis is performed on these components. A Fast Fourier Transform (FFT) is used to convert the time-domain signal into a frequency-domain signal, and the power spectral density of each strain gauge 8's dynamic component is calculated. This density reflects the proportion of vibration energy at different frequencies. From the power spectral density of each strain gauge 8, the frequency component with the highest power value is selected as the local dominant frequency of that individual strain gauge 8. The local dominant frequencies of all strain gauges 8 are statistically averaged to obtain the dominant frequency corresponding to the vibration energy in a specific frequency band for the entire sensor network. Dominant frequency data at different time points are continuously collected. The shift and rate of change of the dominant frequency over time are calculated to form a vibration dominant frequency migration trend. This trend characterizes the evolution of the pipe's vibration spectrum; that is, when structural degradation occurs in the pipe, changes in stiffness lead to a regular shift in the vibration dominant frequency, and the migration amplitude is positively correlated with the degree of degradation.

[0073] The correlation between the vibration dominant frequency migration trend and the change trend of the uniformity of pipe stiffness distribution is verified, and a trend coordination coefficient is calculated. The trend coordination coefficient reflects the synchronicity between the vibration dominant frequency migration trend and the change of the uniformity of pipe stiffness distribution. When the trend coordination coefficient exceeds a preset coordination threshold, it indicates that the dominant frequency migration and stiffness weakening are linked, characterizing the degradation of pipe structural performance. A corresponding structural performance degradation early warning index is generated, providing a basis for subsequent risk assessment.

[0074] Three types of anomaly judgment thresholds are set, including: normal deformation threshold, preset normal gradient threshold, and preset difference threshold, which correspond to the reasonable range of deformation, deformation gradient, and deformation distribution model, respectively. The thresholds are calibrated based on pipeline design parameters, material properties, and historical monitoring data.

[0075] If the deformation corresponding to each spatial coordinate in the overall deformation distribution field exceeds the normal deformation threshold, or the deformation gradient difference exceeds the preset normal gradient threshold, or the difference between the overall deformation distribution field and the preset normal distribution model exceeds the preset difference threshold, then the overall deformation distribution field is decomposed into a direction vector to obtain a deformation direction vector used to characterize the deformation direction.

[0076] The directional vector decomposition is based on continuous data of the overall deformation distribution field and is achieved through spatial coordinate mapping and vector operations. Specifically, a three-dimensional coordinate system is first established with the PCCP pipeline axis as the axial direction, the circumferential tangent as the transverse direction, and the perpendicular to the pipeline wall as the radial direction. The deformation of each spatial point in the deformation distribution field is then converted into displacement data within the coordinate system. Next, the deformation distribution field is divided into multiple independent analysis sub-regions according to reasonable intervals along the pipeline's axial and circumferential directions, ensuring that each sub-region contains a sufficient amount of effective strain data. The displacement data of all points within each sub-region are extracted and the average value is calculated as the representative deformation displacement of that region. The representative deformation displacement of each sub-region is normalized to eliminate the influence of differences in the magnitude of deformation in different regions, retaining only the directional characteristics to obtain the normalized vector for each sub-region. Finally, the normalized vectors of all sub-regions are aggregated and analyzed. The total vector is calculated through vector superposition, and its direction represents the main deformation direction of the abnormal region. Vectors from sub-regions with significant directional differences are extracted separately as secondary deformation directions, ultimately forming a complete set of deformation direction vectors that comprehensively characterize the spatial orientation of deformation.

[0077] Statistical analysis is performed on the deformation direction vector to calculate the main force direction of the pipe body; spatial attenuation fitting is performed on the overall deformation distribution field to obtain the deformation attenuation coefficient, and the stress transmission path is determined based on the deformation attenuation coefficient.

[0078] When calculating the main force direction of the pipe, the deformation direction vectors of all sub-regions are first collected. A vector clustering algorithm is then used to classify vectors with similar directions, and the proportion and aggregation strength of each type of vector are statistically analyzed. The direction corresponding to the vector category with the highest proportion and the strongest aggregation strength is selected as the core candidate direction. Then, combined with the deformation weights corresponding to each vector, a weighted summation is performed to obtain the final main force direction of the pipe. This direction can accurately reflect the core orientation of the external forces acting on the pipe.

[0079] When determining the stress transmission path, multiple characteristic lines are first selected along the axial and circumferential directions of the pipe, and deformation data at different positions on each characteristic line are extracted. Spatial attenuation fitting is performed on the deformation of each characteristic line, and an exponential attenuation model is used to construct a functional relationship between the deformation and the propagation distance. The attenuation coefficient obtained from the fitting can characterize the rate of energy loss along the propagation path; that is, the smaller the attenuation coefficient, the farther the stress is transmitted and the smaller the energy loss; conversely, the transmission distance is short and the energy loss is fast.

[0080] By combining the distribution patterns of deformation attenuation coefficients of all characteristic lines, and tracing the regions where the attenuation coefficient gradient changes smoothly, this region represents the main stress transmission channel. Simultaneously, by referencing the main force directions, the transmission channels of each characteristic line are integrated to form a complete stress transmission path, clearly presenting the stress propagation trajectory and influence range within the tube.

[0081] The energy time history of the characteristic frequency band of the optical fiber sensor is extracted, and the propagation velocity is fitted to the energy time history of the characteristic frequency band. First, the timestamps corresponding to the energy peaks in the energy time history of the characteristic frequency band are extracted. Combined with the spatial resolution of the optical fiber sensor, the specific spatial coordinates of each energy peak in the pipe axis are located. The energy propagation time sequence and spatial distribution relationship are constructed with the timestamp as the horizontal axis and the spatial coordinates as the vertical axis. Linear fitting is performed on this relationship to obtain the energy propagation velocity fitting curve. The slope of the curve is the propagation velocity of the vibration event. If the spatial coordinates increase and the timestamps increase synchronously along a certain axis of the pipe, the propagation direction is consistent with that axis; otherwise, it is the opposite. Combined with the energy distribution gradient of the entire optical fiber, the consistency of the propagation direction is verified, and the propagation direction of the vibration event is finally accurately determined, providing a basis for subsequent disturbance source determination. The angle between the main force direction and the vibration event propagation direction is calculated to obtain the directional consistency coefficient used to characterize the consistency of the two directions. The closer the directional consistency coefficient is to 1, the more consistent the two directions are, and the more likely it is to originate from a directional external disturbance.

[0082] When the directional consistency coefficient exceeds the preset directional threshold and the duration exceeds the preset duration threshold (set according to the typical duration of mechanical operation), the abnormality is determined to originate from directional external disturbance. Combining the abnormal area of ​​the overall deformation distribution field with the propagation range of the vibration event, an early warning information containing the disturbance location and affected area is generated and pushed to external management personnel to prompt them to conduct on-site verification of the area and confirm the legality of the operation and the degree of impact.

[0083] When the anomaly is determined to originate from a directional external disturbance, the deformation amount exceeding the normal deformation threshold and the corresponding deformation gradient difference and deformation attenuation coefficient in the overall deformation distribution field are extracted, and a spatiotemporal sequence of deformation field disturbance is formed by time series processing.

[0084] The initial electromagnetic signals of each magnetic sensor 7 in the sensor network are acquired, and the time-varying data of the amplitude attenuation ratio and phase shift of the initial electromagnetic signals are calculated to form an electromagnetic signal perturbation sequence. The timestamps of the deformation field perturbation sequence and the electromagnetic signal perturbation sequence are compared to obtain the occurrence sequence difference and recovery lag time. Among them, the occurrence sequence difference reflects the order of deformation triggered by the perturbation and the change of electromagnetic signals, and the recovery lag time reflects the synchronicity of deformation and electromagnetic signal recovery.

[0085] The spatial propagation distance and speed of the deformation field disturbance sequence and the electromagnetic signal disturbance sequence along the pipeline axis are analyzed to extract spatial coupling characteristics, including the consistency of propagation speed and the overlap of influence range, and to quantify the spatial correlation between deformation and electromagnetic signals.

[0086] Based on multiple on-site verifications of mechanical operation scenarios, several sets of data on occurrence sequence differences, recovery lag times, and spatial coupling characteristics were collected. Simultaneously, the intensity and duration of external construction loads and the actual damage state of the prestressed steel wire 3 were recorded to construct a training dataset. A random forest model was selected as the perturbation impact assessment model. This model possesses strong anti-interference capabilities and nonlinear fitting ability, and can effectively handle multi-dimensional perturbation characteristic data.

[0087] The training dataset was divided into training and testing sets in a 7:3 ratio and input into the random forest model for training. By adjusting hyperparameters such as the number and depth of decision trees, the model's prediction accuracy was optimized. The model was able to quantify the intensity and duration of external construction loads based on the temporal and spatial characteristics of the input, and then dynamically predict the fracture risk level of prestressed steel wire 3 (divided into four levels: low, medium, high, and extremely high).

[0088] In routine monitoring, the currently acquired sequence difference, recovery lag time, and spatial coupling characteristics are input into the trained disturbance impact assessment model, which outputs the corresponding fracture risk level. If the level is medium or above, a fracture warning is immediately issued, specifying the warning level, affected area, and recommended response measures; if the risk is low, the data change trend is continuously monitored to form a dynamic warning closed loop.

[0089] In practice, the deformation characteristics and the effects of mechanical operations are analyzed through resistance signals to achieve structural degradation assessment and fracture risk warning. The specific implementation is as follows: First, the resistance signals of all strain gauges 8 in the sensor network are collected. A continuous overall deformation distribution field is generated using the Kriging space interpolation algorithm. The deformation gradient difference in the vertical height direction of the pipe is calculated to evaluate the longitudinal uniformity of the pipe stiffness. The dynamic variation components of strain gauge 8 are extracted, and the local dominant frequency of each gauge is obtained through spectrum analysis. The overall dominant frequency is obtained by statistical averaging. The vibration dominant frequency migration trend is continuously monitored and correlated with the stiffness uniformity change trend to calculate the synergy coefficient. If the threshold is exceeded, a structural degradation early warning index is generated.

[0090] Three types of anomaly thresholds are set. If the deformation, gradient difference, or distribution model deviation exceeds the threshold, the deformation field is decomposed into direction vectors. Specifically, a three-dimensional coordinate system is first established, and the analysis sub-regions are divided. After normalization and vector aggregation, the deformation direction vector is obtained. The main force direction of the tube is determined by combining vector clustering and weighted summation. The deformation attenuation coefficient is obtained by fitting characteristic lines, and the stress transmission path is determined by tracking the flat areas and referring to the force direction.

[0091] The energy time history of the characteristic frequency band of the optical fiber sensor is extracted, and the spatiotemporal coordinates of the energy peak are located by combining the spatial resolution of the optical fiber. The propagation speed and direction are obtained by linear fitting. The consistency coefficient between the force direction and the vibration propagation direction is calculated. If the coefficient exceeds the threshold and remains within the standard, it is judged as a directional external disturbance, and an early warning is pushed to prompt management personnel to conduct on-site verification.

[0092] Deformation field and electromagnetic signal disturbance sequences under disturbance scenarios are collected, and the occurrence sequence difference, recovery lag time, and spatial coupling features are extracted to construct a training dataset. A random forest model is used to train the dataset by proportionally dividing it, and the impact of construction loads is quantified after optimizing hyperparameters to dynamically predict the fracture risk level. Real-time feature data is input during daily monitoring, and corresponding early warnings are initiated based on the risk level output by the model. If the risk is low, continuous tracking is performed to form a closed loop.

[0093] The above are merely embodiments of the present invention. The invention is not limited to the fields covered by these embodiments. Commonly known structures and characteristics in the solutions are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are able to access all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for acquiring and analyzing PCCP prestressed failure signals, characterized in that, Includes the following steps: An excitation coil and a magnetic sensor are set on the outside of the prestressed steel wire layer of the PCCP pipe. The excitation coil is used to apply a low-frequency alternating excitation magnetic field to the prestressed steel wire layer, and the magnetic sensor is used to collect the eddy current magnetic field signal induced in the prestressed steel wire to obtain the initial electromagnetic signal of each measuring point. A strain gauge is fixedly connected between the excitation coil and the magnetic sensor. The resistance signal of the strain gauge is acquired in real time, and state characteristic values ​​reflecting the mechanical coupling state of the monitoring layer are extracted from the resistance signal. The state characteristic values ​​include at least the baseline drift rate, vibration energy in a specific frequency band, and the count of sudden jump events. An acoustic fiber optic sensor is laid inside the PCCP pipe along the pipe axis to collect the original vibration signal of the entire pipe length in real time. The amplitude attenuation, phase shift, and frequency response variation characteristics of the initial electromagnetic signal are analyzed and extracted. The first wire breakage range and the corresponding electromagnetic confidence level are calculated. The time-frequency characteristics of the original vibration signal and the correlation between the original vibration signal and the state characteristic value are analyzed. The second wire breakage range and the corresponding acoustic confidence level are calculated. When both electromagnetic confidence and acoustic confidence are higher than the preset confidence threshold, and the first and second wire breakage ranges overlap spatially, the overlapping area is determined as a reliable wire breakage location, and the reliable wire breakage location is used as an alarm output; when there is a numerical difference between electromagnetic confidence and acoustic confidence or the spatial ranges do not overlap, the electromagnetic confidence or acoustic confidence is marked in the first and second wire breakage ranges, and the first and second wire breakage ranges are merged as a contradictory alarm output; When only electromagnetic confidence or acoustic confidence is obtained, and the duration of the electromagnetic confidence or acoustic confidence being higher than the preset confidence threshold exceeds the preset judgment threshold, the state feature value is combined to make a judgment and generate a self-diagnostic prompt.

2. The PCCP prestressing failure signal acquisition and analysis method according to claim 1, characterized in that: The excitation coil is mechanically connected to multiple magnetic sensors through strain gauges, and adjacent magnetic sensors are mechanically connected through strain gauges, thus forming a sensor network with magnetic sensors and excitation coils as nodes and strain gauges as connecting edges.

3. The PCCP prestressing failure signal acquisition and analysis method according to claim 2, characterized in that: When extracting state feature values ​​from resistance signals, the resistance signals of each strain gauge are first filtered to separate the dynamic and slow-changing components. By performing spectral analysis on the dynamic components of all strain gauges within the integrated sensor network, the vibration energy in a specific frequency band reflecting the overall state can be obtained. Calculate the rate of change of the slow component of each strain gauge and statistically analyze the distribution characteristics of the slow component in the sensor network to obtain the baseline drift rate that reflects the overall slow deformation trend of the monitoring layer. A first time window is set. When the number of strain gauges whose dynamic change component amplitude exceeds a preset first amplitude threshold within the first time window exceeds the preset first number threshold, the time corresponding to the first time window and the position of the strain gauge in the sensor network whose dynamic change component amplitude exceeds the preset first amplitude threshold are recorded as a cooperative mutation event. The frequency of occurrence of coordinated mutation events is statistically analyzed, and the spatial distribution correlation between strain gauges in coordinated mutation events is analyzed to obtain a subset of the sudden change event count; A second time window is set. When the amplitude of the dynamic change component of a single strain gauge exceeds a preset second amplitude threshold within the second time window, the time corresponding to the second time window and the position of the strain gauge in the sensor network are recorded as a sudden change event. The second amplitude threshold is greater than the first amplitude threshold. The number of sudden change events occurring within a preset range is counted to obtain another subset of the sudden change event count.

4. The PCCP prestressing failure signal acquisition and analysis method according to claim 2, characterized in that: When analyzing and extracting the characteristics of the initial electromagnetic signal and calculating the electromagnetic confidence score, the initial electromagnetic signals collected by all magnetic sensors in the sensor network are first obtained. The initial electromagnetic signal of each magnetic sensor is compared with the corresponding reference data, and the attenuation ratio, phase shift and frequency response difference are calculated. The initial electromagnetic signals with differences exceeding the preset normal threshold are regarded as abnormal signals. Then, the spatial distribution and clustering pattern of the magnetic sensors that obtained abnormal signals on the sensor network topology are analyzed, and the first wire breakage range and electromagnetic confidence score are calculated based on the spatial distribution and clustering pattern. Among them, the baseline data is the electromagnetic signal baseline data collected by the magnetic sensor when the prestressed steel wire of the PCCP pipe is in an intact and undamaged state.

5. The PCCP prestressing failure signal acquisition and analysis method according to claim 3 or 4, characterized in that: When analyzing the time-frequency characteristics of the original vibration signal and the correlation between the original vibration signal and the state characteristic value, the original vibration signal is first transformed by time and frequency. The sequence of signal energy changing with time within the preset fracture frequency range is extracted as the characteristic frequency band energy time history. The spatiotemporal distribution of the characteristic frequency band energy time history is matched with the spatial distribution of vibration energy in a specific frequency band obtained from the sensor network. The matching degree coefficient is calculated. Combining the intensity, spatiotemporal focus, and matching degree coefficient of the characteristic frequency band energy time history, the range of the second broken wire and the acoustic confidence level are calculated.

6. The PCCP prestressing failure signal acquisition and analysis method according to claim 5, characterized in that: When there are numerical differences or spatial non-overlapping ranges between electromagnetic confidence and acoustic confidence, the dynamic change components of strain gauges corresponding to adjacent magnetic sensors in the sensor network are first obtained; cross-correlation calculation is performed on the dynamic change components of adjacent strain gauges to obtain the correlation coefficient used to characterize the similarity and synchronous change trend of the two dynamic change components. The maximum absolute value or root mean square value of the dynamic change component within the first or second time window is used as the amplitude of the dynamic change component. The correlation coefficient and the amplitude of the strain gauge's dynamic change component are used for time synchronization verification to obtain the disturbance synchronization index of adjacent nodes. Based on the correlation coefficient and synchronization index, a spatial consistency map characterizing the distribution of external vibration interference is constructed. For regions in the first and second broken wire ranges where the synchronization index exceeds a preset synchronization threshold, the weights of electromagnetic confidence and acoustic confidence are reduced to obtain the corrected first and second broken wire ranges. Then, electromagnetic confidence or acoustic confidence is marked in the first and second broken wire ranges.

7. The PCCP prestressing failure signal acquisition and analysis method according to claim 5, characterized in that: When only electromagnetic confidence or acoustic confidence is obtained, and the duration of electromagnetic confidence or acoustic confidence being higher than the preset confidence threshold exceeds the preset judgment threshold, the abnormal signals of all magnetic sensors in the corresponding first or second broken wire range are first extracted, and the attenuation ratio, phase shift and frequency response difference corresponding to the abnormal signal are obtained to obtain the feature difference matrix used to characterize the spatial distribution difference of the abnormal signal. The dynamic variation component amplitudes of adjacent strain gauges are obtained, and correlation analysis is performed in combination with the characteristic difference matrix to obtain the correlation coefficient of abnormal disturbance. Based on the feature difference matrix, it is determined whether the abnormal signal is concentrated in a single magnetic sensor node or is continuously distributed along the sensor network topology. The judgment result is processed into an abnormal location, and then combined with the abnormal disturbance correlation coefficient, a self-diagnostic prompt pointing to the abnormal location is generated.

8. The PCCP prestressing failure signal acquisition and analysis method according to claim 2, characterized in that: First, the resistance signals of all strain gauges in the sensor network are acquired. Then, the resistance signals are processed by a spatial interpolation algorithm to obtain the overall deformation distribution field. Gradient calculations are performed on the overall deformation distribution field to obtain the deformation gradient differences at different locations along the vertical height of the PCCP pipeline. Based on the deformation gradient differences, the longitudinal uniformity of the pipe stiffness distribution is evaluated. The dynamic variation components of all strain gauges in the sensor network are extracted, and the power spectral density of each strain gauge's dynamic component is obtained by spectral analysis. From the power spectral density of each strain gauge, the frequency component with the largest power value is selected as the local dominant frequency of a single strain gauge. The local dominant frequencies of all strain gauges are statistically averaged to obtain the dominant frequency corresponding to the vibration energy of the entire sensor network in a specific frequency band. The dominant frequency data at different time points are continuously collected, and the change of the dominant frequency with time is calculated to obtain the vibration dominant frequency migration trend used to characterize the evolution trend of the vibration spectrum. The correlation between the vibration dominant frequency migration trend and the change trend of pipe stiffness distribution uniformity is verified to obtain the trend coordination coefficient; When the trend synergy coefficient exceeds the preset synergy threshold, a structural performance degradation early warning indicator is generated.

9. The PCCP prestressing failure signal acquisition and analysis method according to claim 8, characterized in that: If the deformation of each spatial coordinate in the overall deformation distribution field exceeds the normal deformation threshold, the deformation gradient difference exceeds the preset normal gradient threshold, or the difference between the overall deformation distribution field and the preset normal distribution model exceeds the preset difference threshold, then the overall deformation distribution field is decomposed into a direction vector to obtain a deformation direction vector used to characterize the deformation direction. Statistical analysis is performed on the deformation direction vector to calculate the main force direction of the tube; spatial attenuation fitting is performed on the overall deformation distribution field to obtain the deformation attenuation coefficient, and the stress transmission path is determined based on the deformation attenuation coefficient. The characteristic frequency band energy time history of the optical fiber sensor is extracted, and the propagation speed is fitted to the characteristic frequency band energy time history to obtain the propagation direction of the vibration event. The angle between the main force direction and the vibration event propagation direction is calculated to obtain the directional consistency coefficient used to characterize the consistency of the two directions. When the directional consistency coefficient exceeds the preset directional threshold and the duration exceeds the preset duration threshold, the anomaly is determined to originate from a directional external disturbance; by combining the anomaly area of ​​the overall deformation distribution field with the propagation range of the vibration event, early warning information including the disturbance location and the affected area is generated.

10. A PCCP prestressed failure signal acquisition and analysis system, characterized in that, The PCCP prestressing failure signal acquisition and analysis method described in any one of claims 1-9 was used.

Images