A built-in self-aware intelligent cable health state monitoring method and system

Abstract

The present application relates to the technical field of material property analysis, and in particular to a built-in self-sensing intelligent cable health state monitoring method and system, which solves the technical problems that in the prior art, stress redistribution caused by internal wire damage of the cable covers up the local damage characteristics, leading to no obvious change in overall stress data, and making it difficult to detect abnormalities and accurately locate the damage position. The method comprises: collecting multi-channel strain data of the cable under load action through a plurality of optical fiber strain sensors arranged in a radial layering and axial segmentation manner inside the cable; identifying an effective load event according to the multi-channel strain data; determining the axial and circumferential positions of the cable damage by analyzing the propagation characteristics of stress waves inside the cable and the strain response characteristics for the effective load event; and determining the radial position of the cable damage and the severity of the damage according to the proportional relationship change between the strain data of different radial layers located at the same axial position.

Description

Technical Field

[0001] This invention relates to the field of materials property analysis technology, specifically to a method and system for monitoring the health status of a built-in self-sensing intelligent cable. Background Technology

[0002] As the main load-bearing components of a cable-stayed bridge, the cables bear complex dynamic forces such as vehicle loads and wind loads over a long period of time. At the same time, they are also exposed to harsh environmental conditions such as rainwater erosion and chloride ion corrosion, making them highly susceptible to damage such as wire corrosion, fatigue breakage, and anchorage degradation. Their health status is directly related to the safe operation of the bridge.

[0003] To monitor the health status of cables, existing technologies are mainly divided into two categories: external sensor methods and internal sensor methods. External sensors are susceptible to environmental corrosion and wind load interference, which can lead to data distortion. Although internal sensor methods can avoid the influence of the external environment and are close to the core stress area, they mostly use discrete sensing units, resulting in monitoring blind spots and making it difficult to achieve full-dimensional coverage monitoring.

[0004] In actual monitoring, the built-in sensing method also faces key technical bottlenecks: when one or more steel wires inside the cable rust or break, stress redistribution will occur between the wire bundles. The load of the damaged wire will be directly transferred to the adjacent intact wires through contact friction and grouting material. As a result, the changes in the macroscopic stress data collected by the sensor are not obvious, and even in the early stage of damage, almost no abnormal signals can be observed. It is difficult to detect local damage in time. Often, by the time the monitoring data shows obvious fluctuations, the damage has already spread to a more serious extent. Summary of the Invention

[0005] To address the technical problem in existing technologies where stress redistribution caused by damage to the internal steel wires of the cable masks localized damage characteristics, resulting in inconspicuous changes in overall stress data and difficulty in detecting anomalies and accurately locating damage, the present invention aims to provide a built-in self-sensing intelligent cable health status monitoring method and system. The specific technical solution adopted is as follows: Firstly, a built-in self-sensing intelligent cable health status monitoring method is provided, comprising: acquiring multi-channel strain data of the cable under load by using multiple fiber optic strain sensors arranged radially in layers and axially in segments inside the cable; identifying effective load events based on the multi-channel strain data; determining the axial and circumferential locations of cable damage for effective load events by analyzing the propagation characteristics of stress waves and strain response characteristics inside the cable; and determining the radial location and severity of cable damage based on the proportional relationship between strain data of different radial layers located at the same axial position.

[0006] Based on the above technical solution, in the built-in self-sensing intelligent cable health status monitoring method provided by this invention, multi-channel strain data is collected by arranging fiber optic strain sensors radially layered and axially segmented inside the cable. Effective load events are accurately identified to eliminate interference. Combining stress wave propagation characteristics and strain response characteristics, the axial and circumferential positions of damage are located. The radial position and severity of damage are determined by relying on the proportional relationship changes of strain data in different radial layers. This effectively overcomes the problem that early local damage caused by stress redistribution due to damage to the steel wires inside the cable is difficult to identify in existing built-in sensing methods. It realizes accurate monitoring of the three-dimensional spatial position of early local damage inside the cable, improves the reliability and pertinence of cable health status monitoring, and provides an accurate basis for targeted maintenance.

[0007] In conjunction with the first aspect above, in one possible implementation, the method of arranging multiple fiber optic strain sensors radially in layers and axially in segments inside the cable specifically includes: arranging a first set of sensors in the gaps between the central strands of the cable to form an inner sensing network; arranging a second set of sensors circumferentially and uniformly in the gaps between the outer strands of the cable to form a middle sensing network; arranging a third set of sensors between the outermost strand and the sheath of the cable to form an outer sensing network; and arranging sensors at different intervals along the cable axis at the anchorage end, transition section, and free section; the sensor spacing at the anchorage end is smaller than the sensor spacing in the transition section and the free section.

[0008] In conjunction with the first aspect above, in one possible implementation, the method for identifying effective load events specifically includes: extracting the instantaneous strain increment of the sensor using a sliding time window of a preset time length; determining the load transfer relationship between multiple sensor measuring points based on the strain increment between different sensor measuring points within the same time window, and constructing a topological network for load transfer; determining a comprehensive index reflecting the stability of the topological network and the degree of load transfer activity based on the load transfer relationship; and determining that if the comprehensive index is greater than a preset threshold, the time period corresponding to the time window is considered an effective load event.

[0009] In conjunction with the first aspect above, in one possible implementation, the method for determining the load transfer relationship between multiple sensor measuring points specifically includes: treating the sensor measuring points as nodes in a mechanical network, determining the load transfer coefficient from one sensor measuring point to the next sensor measuring point based on the strain increment of one sensor measuring point and the strain increment of another sensor measuring point; and forming a load transfer relationship matrix from the load transfer coefficients between multiple sensor measuring points.

[0010] In conjunction with the first aspect above, in one possible implementation, the method for determining the axial and circumferential locations of cable damage by analyzing the propagation characteristics and strain response features of stress waves within the cable under a load event specifically includes: obtaining the actual arrival time of the stress wave from the excitation point to each sensor measuring point under a load event; determining the theoretical arrival time based on the spatial location of each sensor measuring point; and selecting sensor measuring points with abnormal arrival time delays based on the difference between the actual arrival time and the theoretical arrival time of each sensor measuring point, thereby determining the axial and circumferential locations of the cable damage.

[0011] In conjunction with the first aspect above, in one possible implementation, the method for determining the axial and circumferential locations of cable damage specifically includes: extracting the response curves of sensor measuring points selected as having anomalies in arrival time delay during the strain rise phase; analyzing the variation characteristics of the response curves to determine a path complexity index characterizing the complexity of the stress transmission path; and determining the axial and circumferential locations of cable damage based on the sensor measuring points with anomalies in arrival time delay and the path complexity index.

[0012] In conjunction with the first aspect above, in one possible implementation, the method for determining the radial location and severity of cable damage based on the proportional relationship between strain data of different radial layers located at the same axial position specifically includes: extracting the strain increments of inner, middle, and outer layer measuring points at the same axial position during one load cycle; constructing a two-dimensional phase space based on a first ratio of the strain increment of the middle layer measuring point to the strain increment of the inner layer measuring point, and a second ratio of the strain increment of the outer layer measuring point to the strain increment of the middle layer measuring point; and determining the radial layer where the damage occurred and assessing the degree of damage based on the offset direction and distance of the monitoring data points relative to the healthy data point cluster in the two-dimensional phase space.

[0013] In conjunction with the first aspect above, in one possible implementation, the method for determining the radial layer where damage occurs based on the offset direction and distance of the monitoring data points relative to the health data point cluster specifically includes: if the first proportion value of the monitoring data points is greater than the first proportion value of the center of the health data point cluster, and the difference between the second proportion value and the second proportion value of the health data point cluster center is within a preset range, then the damage is determined to occur in the inner layer; if the second proportion value of the monitoring data points is less than the second proportion value of the center of the health data point cluster, and the difference between the first proportion value and the first proportion value of the health data point cluster center is within a preset range, then the damage is determined to occur in the outer layer; if the first proportion value of the monitoring data points is greater than the first proportion value of the center of the health data point cluster, and the second proportion value of the monitoring data points is less than the second proportion value of the center of the health data point cluster, then the damage is determined to occur in the middle layer.

[0014] In conjunction with the first aspect above, in one possible implementation, the method for determining the health data point cluster specifically includes: under the cable health state, collecting first and second proportional values ​​corresponding to multiple complete load events, and forming a data point distribution in a two-dimensional phase space; using a preset density estimation algorithm to analyze the data point distribution, and determining the core region and boundary of the health data point cluster.

[0015] Secondly, a built-in self-sensing intelligent cable health status monitoring system is provided, comprising: a layered built-in sensor network module, a data demodulation and acquisition module, and a data processing and analysis module; the layered built-in sensor network module includes multiple fiber optic strain sensors arranged radially in layers and axially in segments inside the cable, used to collect multi-channel strain data of the cable under load; the data demodulation and acquisition module is connected to the layered built-in sensor network module, used to demodulate and acquire the strain data of the fiber optic strain sensors in real time; the data processing and analysis module is communicatively connected to the data demodulation and acquisition module and is configured to perform: identifying effective load events from the multi-channel strain data; determining the axial and circumferential locations of cable damage by analyzing the propagation characteristics of stress waves and strain response characteristics inside the cable for effective load events; and determining the radial location and severity of cable damage based on the proportional relationship between strain data of different radial layers located at the same axial position.

[0016] The present invention has the following beneficial effects: By collecting multi-channel strain data through radially layered and axially segmented fiber optic strain sensors inside the cable, effective load events can be accurately identified to eliminate interference. Combining stress wave propagation characteristics and strain response features, the axial and circumferential locations of damage can be located. The radial location and severity of damage can be determined based on the proportional changes in strain data from different radial layers. This effectively overcomes the problem of difficulty in identifying early local damage caused by stress redistribution due to damage to the steel wires inside the cable in existing built-in sensing methods. It achieves accurate three-dimensional spatial monitoring of early local damage inside the cable, improves the reliability and targeting of cable health status monitoring, and provides an accurate basis for targeted maintenance. Attached Figure Description

[0017] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a system structure diagram of a built-in self-sensing intelligent cable health status monitoring system provided in one embodiment of the present invention; Figure 2 A flowchart illustrating a method for monitoring the health status of a built-in self-sensing intelligent cable, as provided in one embodiment of the present invention; Figure 3 A schematic diagram of a radially layered fiber optic strain sensor provided in one embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of a built-in self-sensing intelligent cable health status monitoring device according to an embodiment of the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a built-in self-sensing intelligent cable health status monitoring method and system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0021] The following description, in conjunction with the accompanying drawings, details the specific scheme of the built-in self-sensing intelligent cable health status monitoring method and system provided by the present invention.

[0022] Please see Figure 1 The diagram illustrates a system structure of a built-in self-sensing intelligent cable health status monitoring system according to an embodiment of the present invention. The built-in self-sensing intelligent cable health status monitoring system includes: a hierarchical built-in sensor network module 1, a data demodulation and acquisition module 2, and a data processing and analysis module 3.

[0023] The core foundation of the layered built-in sensor network module 1 system for data acquisition is a multi-dimensional monitoring network constructed by synchronously embedding fiber optic strain sensors inside the cable. The acquired multi-channel strain data directly provides the raw data source for subsequent modules. The module comprises three sub-modules, all installed during the cable wire bundle stranding stage. They are arranged utilizing the natural gaps between the wire bundles, do not replace the cable wires, and are fixed before the secondary grouting or extrusion of the polyethylene (PE) sheath. The inner sensing submodule 11 consists of fiber optic strain sensors corresponding to 2-3 optical fibers, which are arranged in the gaps of the outer steel wires of the central strand of the cable, and can accurately capture the strain state of the core load-bearing area of ​​the cable. The middle layer sensing submodule 12 consists of fiber optic strain sensors corresponding to 3-4 optical fibers, which are uniformly arranged circumferentially in the gaps between the outer strands of steel wires, with circumferential intervals of 90° or 120°, and can monitor strain differences in different circumferential directions. The outer sensing submodule 13 consists of fiber optic strain sensors corresponding to 1-2 optical fibers, which are arranged between the outermost steel wire of the cable and the PE sheath. It is highly sensitive to damage such as water seepage into the sheath and corrosion of the steel wire surface.

[0024] The sensors of the three sub-modules are arranged in sections along the cable axis: anchor end, transition section, and free section. The sensors are most closely spaced at the anchor end, the spacing is moderately widened in the transition section, and they are arranged at equal intervals in the free section to ensure that there are no blind spots in the monitoring along the entire length of the cable. The strain signals collected by all sensors are collected and transmitted through the fiber optic trunk line.

[0025] The data demodulation and acquisition module 2 takes over the signal transmission tasks from the hierarchical built-in sensor network module 1, and is responsible for converting the optical signals acquired by the sensors into analyzable electrical signals, forming standardized data for subsequent modules to process. The module contains two functional sub-modules: The data transmission submodule 21 uses an optical fiber trunk spirally wound around the outside of the steel wire bundle along the cable axis as a physical transmission carrier to connect all the optical fiber strain sensors of the inner, middle and outer sensing submodules in series and lead them out at the end of the cable to achieve low-loss and anti-interference transmission of strain signals. The core physical device of the data demodulation submodule 22 is a wavelength scanning fiber optic grating demodulator. By continuously scanning the entire fiber optic link, it calculates the strain value of each fiber optic strain sensor in real time and converts the optical signal into a multi-channel strain time history data stream. This data stream is transmitted in real time to the next stage data processing and analysis module 3 to provide standardized data support for damage identification.

[0026] The data processing and analysis module 3 is the core computing unit of the system. Based on the multi-channel strain time history data stream output by the data demodulation and acquisition module 2, it uses a series of intelligent algorithms to achieve precise damage location and severity assessment. The processing results directly provide decision-making basis for cable maintenance. The module contains three core sub-modules: The effective load event identification submodule 31 extracts the instantaneous strain increment through a sliding time window of a preset time length, regards the steel wire bundle in the cable section as a mechanical network node, determines the load transfer relationship based on the strain increment of different sensor measurement points, constructs a load transfer topology network, calculates a comprehensive index reflecting the network stability and load transfer activity, and filters out effective load events after comparing with a preset threshold. The filtered effective strain data will be synchronously transmitted to the subsequent two submodules. The axial and circumferential damage location submodule 32 receives the effective strain data output by the effective load event identification submodule 31, obtains the actual arrival time of the stress wave from the excitation point to each sensor measurement point, determines the theoretical arrival time by combining the spatial location of the measurement point, filters out measurement points with abnormal arrival time delay by analyzing the difference between the two, extracts the response curve of the strain rise stage of these measurement points, analyzes the curve change characteristics to determine the path complexity index, and finally locks the axial and circumferential location of the cable damage by combining the abnormal delay situation. The radial damage and severity determination submodule 33 extracts the strain increments of the inner, middle, and outer layer measuring points at the same axial position, calculates the first ratio of the strain increments of the middle layer to the inner layer and the second ratio of the strain increments of the outer layer to the middle layer, constructs a two-dimensional phase space using the two as coordinates, analyzes the core area and boundary of the healthy data point cluster under the healthy state of the cable through a preset density estimation algorithm, and then determines the radial layer where the damage occurred based on the offset direction and distance of the monitoring data points relative to the healthy data point cluster, and assesses the severity of the damage.

[0027] Please see Figure 2 The diagram illustrates a flowchart of a method for monitoring the health status of a built-in self-sensing intelligent cable according to an embodiment of the present invention. This method includes: S1. Multiple fiber optic strain sensors are arranged radially in layers and axially in segments inside the cable to collect multi-channel strain data of the cable under load.

[0028] In some implementations, such as Figure 3 As shown, sensors are arranged radially in layers to construct a full-section monitoring network: the first set of sensors is arranged in the gaps between the central strands of the cable to form an inner sensor network; the second set of sensors is evenly arranged circumferentially in the gaps between the outer strands of the cable to form a middle sensor network; and the third set of sensors is arranged between the outermost strands of the cable and the sheath to form an outer sensor network.

[0029] Sensors are arranged in segments along the axial direction to cover the entire length of the monitoring range: The cable is divided into three regions along its axis: the anchorage end, the transition section, and the free section. Sensors are arranged at different intervals along the cable's axis in these three sections. The sensor spacing at the anchorage end is smaller than that in the transition and free sections. Specifically, the anchorage end is an area with significant stress concentration, and a denser spacing allows for accurate capture of stress changes in this area. The stress state in the transition section is relatively mild, and a wider spacing balances monitoring accuracy and cost. Equal spacing in the free section ensures full coverage monitoring of the entire cable length, avoiding blind spots. For example, the spacing at the anchorage end is set to 5cm, the spacing in the transition section to 10cm, and the spacing in the free section to 20cm.

[0030] Furthermore, during the cable assembly process, the fiber optic trunk, inscribed with multiple fiber Bragg grating sensors, is fixed to the gaps or surfaces of the corresponding layers of steel wire bundles in a helical winding manner according to pre-designed radial and axial coordinates. This ultimately forms a sensor network that is physically helically arranged and logically interconnected by all sensors through the same fiber. The fiber optic trunk is led out at the cable end and connected to a wavelength-scanning fiber Bragg grating demodulator. During system initialization, the precise three-dimensional spatial coordinates of each fiber Bragg grating sensor on the helical path are measured or calibrated and recorded. The demodulator is then started to continuously scan the entire fiber optic link. Based on the calibrated spatial coordinates, the calculated strain values ​​can be mapped one-to-one with the specific spatial locations of the sensors, thus forming a multi-channel strain time-history data stream with clear spatial attributes. This helical winding arrangement facilitates strain coordination between the fiber and the steel wire and enables low-loss, interference-resistant signal transmission, providing reliable raw data for subsequent analysis.

[0031] S2. Identify effective load events based on multi-channel strain data.

[0032] In some implementations, the method of S2 can be specifically implemented through the following S21 to S24, which are explained in detail below: S21. Extract the instantaneous strain increment of the sensor using a sliding time window of a preset time length.

[0033] In some implementations, considering typical load scenarios of cable-stayed bridge cables (e.g., when a vehicle passes at 60-80 km / h, the complete loading-unloading cycle of a single vehicle load on the cable is 1.5-3 seconds), a sliding time window with a preset length of 2 seconds is selected, and the sliding step size is set to 0.1 seconds. Based on this window and step size, the multi-channel strain time history data is extracted segment by segment to obtain the instantaneous strain data corresponding to each window. The strain increment of each sensor measuring point within the same window (i.e., the difference between the strain value at the current moment and the strain value at the beginning of the window) is calculated, eliminating the influence of the strain reference value and focusing on the strain change characteristics under load.

[0034] S22. Based on the strain increments between different sensor measuring points within the same time window, determine the load transfer relationship between multiple sensor measuring points and construct a load transfer topology network.

[0035] Unless otherwise specified, all strain values ​​and strain increments mentioned below are in microstrain (µs). ).

[0036] In some implementations, the method for determining the load transfer relationship between multiple sensor measuring points specifically includes: treating the sensor measuring points as nodes in a mechanical network, and determining the load transfer coefficient from one sensor measuring point to the next based on the strain increment of one sensor measuring point and the strain increment of another. Then, a load transfer relationship matrix is ​​formed from the load transfer coefficients among the multiple sensor measuring points. The above process specifically includes: First, all sensor measurement points are classified and numbered radially. For example, measurement points in the inner layer sensor network are numbered 1-4, those in the middle layer sensor network are numbered 5-7, and those in the outer layer sensor network are numbered 8-9. Simultaneously, each numbered measurement point is considered a node in the mechanical network within the cable cross-section. Each node corresponds to a strain monitoring point at a specific spatial location on the cable. This provides a unified identification system for subsequently quantifying the load transfer relationship between measurement points and clarifies the spatial correspondence of each monitoring point.

[0037] Then, using the strain increment at the sensor measurement point as input, the load transfer coefficient between any two measurement points is calculated: In the formula, This represents the load transfer coefficient from measuring point i to measuring point j, used to characterize the degree of influence of strain change at measuring point i on strain change at measuring point j. This represents the strain increment at measuring point i within the current time window, characterizing the intensity of load change at the location of measuring point i. This represents the strain increment at measuring point j within the same time window, characterizing the load variation intensity at the location of measuring point j; α is the parameter adjustment coefficient, if... When α is 0, the value of α is set to 1.0. When the denominator is not zero, α is set to 0 to avoid the calculation being meaningless due to a denominator of 0; The ratio of the strain increment at measuring point j to the strain increment at measuring point i is used to quantify the load transfer intensity from measuring point i to measuring point j. The larger the ratio, the stronger the load transfer effect from i to j.

[0038] Finally, the numbers of all measuring points are used as rows and columns of a matrix (rows represent input measuring points, columns represent output measuring points), and the calculated load transfer coefficients are filled into the corresponding positions. Simultaneously, the elements on the diagonal of the matrix (i.e., the transfer coefficient from the measuring point itself to itself) are set to 1. As shown in Table 1, inner measuring point 1 corresponds to row 1 and column 1 of the matrix, and its diagonal elements ( , The load transfer factor for (etc.) is 1.

[0039] Table 1 By integrating the dispersed load transfer coefficients into a structured transfer matrix, a load transfer topology network within the cable section is constructed, which intuitively presents the load transfer relationship between each measuring point.

[0040] S23. Based on the load transfer relationship, determine a comprehensive index that reflects the stability of the topological network and the degree of load transfer activity.

[0041] In some implementations, the first step is to calculate the topological stability index: First, calculate the character determination factor: In the formula, The role determination factor of measurement point i is used to characterize the functional positioning of the measurement point in the topology network; It represents the total output capacity of measuring point i, which is the sum of all the transfer coefficients in the row corresponding to measuring point i in the transfer matrix, and characterizes the total strength of the load transferred from measuring point i to other measuring points. The total input influence of measuring point i is represented by the sum of all transfer coefficients in the column corresponding to measuring point i in the transfer matrix, characterizing the total intensity of loads received by measuring point i from other measuring points; β is the parameter tuning coefficient, if... When β is 0, the value of β is set to 0.001. When the denominator is not zero, β is set to 0 to avoid the denominator being zero; The ratio of total output capacity to total input impact can determine the role of a measuring point in the network. The larger the ratio, the more inclined the measuring point is to transmit loads to other measuring points, and vice versa.

[0042] Then, the coefficient of variation (CV) of the role determination factors for all measuring points is calculated to characterize the degree of distribution fluctuation of the role determination factors for each measuring point. It should be noted that since the load transfer coefficient can be zero, the corresponding role determination factor can also be zero. To avoid the denominator being zero, a parameter adjustment coefficient is added to the denominator of the general coefficient of variation formula. If the average value is 0, the parameter adjustment coefficient is set to 0.001; if the average value is not 0, the parameter adjustment coefficient is set to 0.

[0043] Next, the topology stability index is obtained through the reciprocal transformation. The smaller the coefficient of variation, the more consistent the role distribution; this is transformed into a larger topology stability index, indicating a more stable network, which is positively correlated with stability. It's important to note that since the standard deviation of the role determination factors for all measurement points can be zero, the corresponding coefficient of variation can also be zero. To avoid the denominator being zero, a parameter tuning coefficient is added to the denominator of the reciprocal transformation formula. If the coefficient of variation is 0, the parameter tuning coefficient is set to 0.001; if the coefficient of variation is not zero, the parameter tuning coefficient is set to 0.

[0044] The second step is to calculate the load transfer activity index: In the formula, S represents the load transfer activity index, used to characterize the load transfer intensity between different measurement points in the topological network; K represents the total number of off-diagonal elements in the transfer matrix, i.e., all The number of elements; This represents the absolute value of the load transfer coefficient from measuring point i to j, eliminating the influence of positive and negative directions; The arithmetic mean of the absolute values ​​of all off-diagonal transfer coefficients is used to quantify the load transfer activity across measuring points in the entire network; the larger the average value, the more frequent the load transfer between different measuring points.

[0045] The third step is to calculate the comprehensive index Z: In the formula, the difference in magnitude is first eliminated by a normalization function. The topological stability index C and the load transfer activity index S are necessary coexistence conditions for effective load events: an effective load event needs to simultaneously satisfy topological stability (clear positioning of measuring point roles) and load transfer activity (significant strain correlation between measuring points). Neither can be lacking. If only S is high but C is low (such as in a noisy scenario: large strain fluctuations but chaotic measuring point roles), or only C is high but S is low (such as in a static scenario: clear measuring point roles but no load transfer), it does not belong to an effective load event. Therefore, multiplication calculation is used to obtain a comprehensive index that can accurately distinguish between effective load and noise. Under real load action, the load transfer network of the cable is both stable (clear roles) and active (frequent transfer), and the comprehensive index will be significantly higher than that of the noisy scenario.

[0046] Unless otherwise specified, the normalization function norm(*) mentioned in the embodiments of this invention uses maximum and minimum value normalization. The maximum and minimum values ​​are preset empirical extreme values ​​derived from a large amount of historical experimental data. If the calculation result exceeds the interval [0, 1], it is restricted to the range [0, 1] by a truncation function (i.e., if the result is less than 0, it is taken as 0; if it is greater than 1, it is taken as 1) to eliminate the influence of outliers on the result.

[0047] S24. If the comprehensive index is greater than the preset threshold, the time period corresponding to the time window is determined to be an effective load event.

[0048] In some implementation methods, a preset threshold is first determined: under the normal traffic (daily average traffic flow) scenario of a cable-stayed bridge, comprehensive indicators of 100 consecutive time windows are collected, and after sorting these indicators from largest to smallest, the top 25th percentile is selected as the benchmark threshold.

[0049] The comprehensive index of the current time window is calculated in real time. If the index is greater than the benchmark threshold, the time period corresponding to the time window is determined to be an effective load event. If the comprehensive index of multiple consecutive time windows is greater than the benchmark threshold, these consecutive windows are merged into a complete effective load event. Effective data segments under actual load are selected from continuous strain data, and invalid data under static conditions or noise interference are excluded.

[0050] S3. For effective load events, the axial and circumferential locations of cable damage are determined by analyzing the propagation characteristics of stress waves inside the cable and the strain response characteristics.

[0051] In some implementations, the method of S3 can be specifically implemented through the following S31 to S33, which are explained in detail below: S31. Obtain the actual arrival time of the stress wave from the excitation point to each sensor measurement point under the effective load event.

[0052] The excitation point is the starting point of the cable stress corresponding to the load change event, such as the cable connection point corresponding to the vehicle load acting on the bridge deck; if the cable is anchored in the middle of the bridge deck, the excitation point is the midpoint of the cable; if the cable is anchored in the bridge tower, the excitation point is the anchorage position of the bridge tower.

[0053] In some implementations, the automatic identification method for excitation points includes: finding the sensor measurement point in the multi-channel strain data where the strain first changes abruptly and the magnitude of the change is the largest; and combining the bridge structure model and vehicle load location information, mapping the measurement point to the corresponding cable spatial location as the excitation point for this load event.

[0054] In some implementations, the multi-channel strain time history data corresponding to the effective load event is retrieved, and the strain time history curve of each sensor measuring point is extracted (i.e., the relationship curve of the strain value collected by the sensor measuring point changing with time; a data point on the curve corresponds to the strain state of the cable wire at the location of the measuring point at a certain moment).

[0055] In an effective load event, the curve will show the process of strain rising as the load is applied and falling as the load is removed. The moment when the strain first begins to rise in each curve is identified and defined as the actual arrival time of the stress wave at the corresponding measuring point.

[0056] S32. Determine the theoretical arrival time based on the spatial location of each sensor measurement point.

[0057] In some implementations, based on the characteristic that stress wave propagation in a healthy cable follows the "shortest distance principle", the geometric distance from each sensor measuring point to the excitation point is first calculated, and then the geometric distance is divided by the equivalent wave velocity that reflects the actual cable structure characteristics (the equivalent wave velocity is obtained through calibration experiments under the cable's healthy state). The result is the theoretical arrival time of that measuring point.

[0058] S33. Based on the difference between the actual arrival time and the theoretical arrival time of each sensor measurement point, filter out the sensor measurement points with abnormal arrival time delays, and determine the axial and circumferential locations of cable damage.

[0059] In some implementations, the difference between the actual arrival time and the theoretical arrival time (i.e., the delay) of each sensor measurement point is calculated. Then, based on the spatial distribution characteristics of the delay, the preliminary suspected damage range is determined in two steps: The first step is to determine the preliminary suspected axial damage area: extract the delay values ​​of all circumferential measuring points at the same axial position and calculate the variance of this set of delay values; select a preset variance threshold (e.g., empirically set to 0.0015s²). If the variance of this set of delay values ​​is greater than the threshold, it indicates that when the stress wave reaches the cross-section, it encounters asymmetrical obstruction on different circumferential paths, and the damage may be located near the cross-section or on one side of the cross-section. This axial position is marked as the suspected axial position. Conversely, when the variance is less than or equal to the preset threshold, it indicates that the delay values ​​of each circumferential measuring point at the same axial position are uniformly distributed. When the stress wave reaches all circumferential paths of the axial cross-section, there is no path difference caused by local damage in the circumferential direction of the axial cross-section, and the damage does not exist near the axial cross-section.

[0060] The second step is to determine the preliminary suspected circumferential damage range: extract the delay values ​​of different circumferential measuring points at the same axial position, and calculate the average delay value of all circumferential measuring points; if the delay value of a certain circumferential measuring point is greater than 25% of this average value, and greater than 10% of the second largest delay value (i.e., the second largest delay value after sorting all circumferential measuring points at the same axial position in descending order of value), this circumferential orientation is marked as the suspected circumferential orientation of that axial position. The 25% and 10% values ​​are empirically calibrated thresholds obtained based on cable health status experiments and damage simulation tests.

[0061] Traverse all axial positions and find spatial points that simultaneously meet the following conditions: the axial position is marked as a suspected axial position (large variance, indicating local asymmetric interference), and the circumferential orientation is marked as a suspected circumferential orientation of the axial position (prominent delay, indicating the most severe path obstruction in that orientation).

[0062] The spatial region corresponding to the sensor measuring point that meets the above conditions is the axial and circumferential location of the cable damage.

[0063] In some implementations, methods for determining the axial and circumferential locations of cable damage specifically include: first, extracting the response curves of sensor measuring points selected as having anomalies in arrival time delay during the strain rise phase; then analyzing the changing characteristics of the response curves to determine a path complexity index characterizing the complexity of the stress transmission path. Specifically, this includes: calculating the second derivative time-series curve of the curve segment; statistically analyzing the number and amplitude of peaks in the second derivative curve, calculating their product, and then performing Z-score normalization on this product in a dataset of calculation results from several healthy load events. The result obtained is the path complexity index. Furthermore, a preset threshold (e.g., 0.3) is selected; if the path complexity index of a measuring point is greater than this threshold, it is marked as a path-complex measuring point.

[0064] Finally, the measurement points with abnormal arrival time delay and the measurement points with complex paths are superimposed. The spatial region corresponding to the measurement point that satisfies both conditions is the axial and circumferential position of the cable damage. By combining the time delay of stress wave propagation and the step characteristics of stress transmission path, the risk of misjudgment by a single indicator is eliminated, and the axial and circumferential positions of the damage are accurately located.

[0065] S4. Based on the proportional relationship between the strain data of different radial layers located at the same axial position, determine the radial location and severity of the cable damage.

[0066] In some implementations, method S4 can be specifically implemented using S41 to S43, which are explained in detail below: S41. Extract the strain increments of the inner, middle and outer measuring points at the same axial position during one load cycle.

[0067] When a single vehicle passes through the bridge deck area corresponding to the cable, the complete process of load loading (vehicle approaching) and load unloading (vehicle leaving) experienced by the cable is called one load cycle.

[0068] In some implementations, the inner, middle, and outer layer measuring points corresponding to the axial location of the damage are extracted. Strain time history data from a single load cycle at that axial location is retrieved. The peak strain time on the strain time history curves of each measuring point within that load cycle is determined. The difference between the strain value at the peak strain time and the strain value at the start of the cycle is calculated and used as the strain amplitude of the measuring point under that load cycle. This amplitude is then used as the strain increment for subsequent calculations. Using the strain amplitude effectively avoids the problem of excessively small increments caused by strain recovery after unloading, making the calculation of strain ratios between different radial layers more stable and significant.

[0069] S42. A two-dimensional phase space is constructed based on the first ratio of the strain increment of the middle layer measuring point to the strain increment of the inner layer measuring point, and the second ratio of the strain increment of the outer layer measuring point to the strain increment of the middle layer measuring point.

[0070] The first and second proportional values ​​utilize the constraint differences between the different radial layers (the inner layer is radially constrained by the outer layer, the outer layer is constrained by the PE sheath, and the middle layer is coupled). It should be noted that regardless of whether the denominator (inner and middle layer strain increments) is zero, a parameter adjustment coefficient of 0.001 is added to the denominator of the calculated proportional value to ensure calculation stability.

[0071] Subsequently, a two-dimensional phase space is established with the first proportional value as the horizontal axis and the second proportional value as the vertical axis. Substituting the first and second proportional values ​​corresponding to the load cycles into the coordinate system, a corresponding data point can be obtained in the two-dimensional phase space. The stable strain proportional relationship under healthy conditions is transformed into spatial coordinates, and the abnormal proportional relationship caused by subsequent damage can be intuitively presented through the offset of the data point.

[0072] S43. In two-dimensional phase space, based on the offset direction and distance of the monitoring data points relative to the cluster of healthy data points, determine the radial layer where the damage occurred and assess the degree of damage.

[0073] In some implementations, under the healthy state of the cable, first and second proportional values ​​corresponding to multiple complete load events are collected, forming a data point distribution in a two-dimensional phase space. Then, a preset density estimation algorithm (such as kernel density estimation (KDE)) is used to analyze the data point distribution to determine the core region and boundaries of the healthy data point clusters. Specifically, this includes: generating a Gaussian kernel function centered on each healthy data point, where the bandwidth of the kernel function is adaptively determined according to the Silverman criterion (this criterion can automatically match an appropriate bandwidth based on the data dispersion, avoiding errors from manual parameter tuning); superimposing the Gaussian kernel functions corresponding to all healthy data points to obtain a density distribution surface in the two-dimensional phase space; extracting the density peak points of the density distribution surface as the center coordinates of the healthy data point clusters; delineating the boundary region of the healthy data point clusters using contour lines where the density value drops to 50% of the peak value. Based on the statistical characteristics of kernel density estimation, the probability density at the Gaussian distribution mean ± 1 standard deviation corresponding to the contour lines where the density value drops to 50% of the peak value can cover approximately 68% of the core distribution of the healthy data, effectively characterizing the inherent dispersion under the healthy state. Meanwhile, the 50% peak criterion can be statistically verified and optimized using pre-collected cable health status samples to ensure that it can stably define the core area of ​​health data point clusters under different load conditions and environmental noise, taking into account both the sensitivity and specificity of damage detection.

[0074] In some implementations, if the first proportion value of the monitored data point is greater than the first proportion value of the health data point cluster center, and the difference between the second proportion value and the second proportion value of the health data point cluster center is within a preset range (the fluctuation range depends on the structural characteristics, such as the absolute value of the difference being less than or equal to 0.1), then the damage is determined to occur in the inner layer. If the second proportional value of the monitoring data point is less than the second proportional value of the health data point cluster center, and the difference between the first proportional value and the first proportional value of the health data point cluster center is within a preset range, then the damage is determined to occur in the outer layer. If the first proportion value of the monitoring data point is greater than the first proportion value of the health data point cluster center, and the second proportion value of the monitoring data point is less than the second proportion value of the health data point cluster center, then the damage is determined to occur in the middle layer.

[0075] Furthermore, the Euclidean distance between the monitoring data points and the center of the healthy data point cluster can be calculated. The greater the distance, the higher the severity of the damage. The angle between the offset direction of the monitoring data points and the coordinate axis can also be calculated to help verify the determination results of the damaged layer.

[0076] The above criteria are based on an ideal model of local stiffness damage occurring in a single radial layer. In actual monitoring, if the offset of the monitoring data points does not meet any of the above typical patterns (for example, the first proportion value is less than the first proportion value of the center of the healthy data point cluster), the system should mark this state as "abnormal unidentified state" or "complex / special damage suspected state". At this time, other auxiliary diagnostic mechanisms such as manual inspection can be triggered.

[0077] Based on the above technical solution, by arranging fiber optic strain sensors radially layered and axially segmented inside the cable to collect multi-channel strain data, effective load events can be accurately identified to eliminate interference. Combining stress wave propagation characteristics and strain response characteristics, the axial and circumferential locations of damage can be located. The radial location and severity of damage can be determined by relying on the proportional relationship changes of strain data in different radial layers. This effectively overcomes the problem that early local damage caused by stress redistribution due to damage to the steel wires inside the cable is difficult to identify in existing built-in sensing methods. It realizes accurate monitoring of the three-dimensional spatial location of early local damage inside the cable, improves the reliability and pertinence of cable health status monitoring, and provides an accurate basis for targeted maintenance.

[0078] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0079] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0080] In this embodiment of the invention, the built-in self-sensing intelligent cable health status monitoring device can be divided into functional units according to the above method example. For example, each function can be divided into its own functional unit, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or as a software functional unit. It should be noted that the unit division in this embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0081] This invention also provides a schematic diagram of the hardware structure of a built-in self-sensing intelligent cable health status monitoring device, see below. Figure 4 The built-in self-sensing smart cable health status monitoring device 400 includes a processor 401, and optionally, a memory 402 connected to the processor 401.

[0082] In the first possible implementation, see Figure 4 The built-in self-sensing intelligent cable health status monitoring device 400 also includes a transceiver 403. The processor 401, memory 402, and transceiver 403 are connected via a bus. The transceiver 403 is used to communicate with other devices or communication networks. Optionally, the transceiver 403 may include a transmitter and a receiver. The device in the transceiver 403 that implements the receiving function can be considered as a receiver, which is used to perform the receiving steps in the embodiments of the present invention. The device in the transceiver 403 that implements the transmitting function can be considered as a transmitter, which is used to perform the transmitting steps in the embodiments of the present invention.

[0083] Based on the first possible implementation method Figure 4 The structural diagram shown can be used to illustrate the structure of the built-in self-sensing intelligent cable health status monitoring device involved in the above embodiments.

[0084] in, Figure 4 The diagram also illustrates the system chip in the built-in self-sensing intelligent cable health status monitoring device. In this case, the actions performed by the aforementioned built-in self-sensing intelligent cable health status monitoring device can be implemented by this system chip; the specific actions performed are described above and will not be repeated here.

[0085] Although the invention has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings and the disclosure, will understand and implement other variations of the disclosed embodiments in carrying out the claimed invention. In this invention, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality. A single processor or other unit can implement several of the functions listed in this invention.

[0086] Although the invention has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made therein without departing from the spirit and scope of the invention. Accordingly, this specification and drawings are merely illustrative of the invention and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if such modifications and modifications of the invention fall within the scope of the invention and its equivalents, the invention is also intended to include such modifications and modifications.

Claims

1. A built-in self-aware intelligent cable health condition monitoring method, characterized in that, include: Multiple fiber optic strain sensors are arranged radially in layers and axially in segments inside the cable to collect multi-channel strain data of the cable under load. Multiple fiber optic strain sensors are arranged radially in layers and axially in segments inside the cable, including: a first set of sensors arranged in the gaps between the central strands of the cable to form an inner sensing network; a second set of sensors evenly arranged circumferentially in the gaps between the outer strands of the cable to form a middle sensing network; a third set of sensors arranged between the outermost strands of the cable and the sheath to form an outer sensing network; and sensors arranged at different intervals along the cable axis at the anchorage end, transition section, and free section; the sensor spacing at the anchorage end is smaller than the sensor spacing in the transition section and the free section. Based on the multi-channel strain data, identify effective load events; For the aforementioned effective load event, the axial and circumferential locations of cable damage are determined by analyzing the propagation characteristics of stress waves inside the cable and the strain response characteristics. The radial location and severity of cable damage are determined based on the proportional relationship between strain data from different radial layers located at the same axial position. This includes: extracting the strain increments of inner, middle, and outer layer measuring points at the same axial position during one load cycle; constructing a two-dimensional phase space based on a first ratio of the strain increment of the middle layer measuring point to the strain increment of the inner layer measuring point, and a second ratio of the strain increment of the outer layer measuring point to the strain increment of the middle layer measuring point; and determining the radial layer where the damage occurred and assessing the degree of damage based on the offset direction and distance of the monitoring data points relative to the healthy data point cluster within the two-dimensional phase space.

2. The built-in self-sensing intelligent cable health status monitoring method according to claim 1, characterized in that, Identify effective load events, including: The instantaneous strain increment of the sensor is extracted using a sliding time window of a preset time length; Based on the strain increments between different sensor measurement points within the same time window, the load transfer relationship between multiple sensor measurement points is determined, and a topology network for load transfer is constructed. Based on the load transfer relationship, a comprehensive index is determined that reflects the stability of the topological network and the degree of load transfer activity. If the comprehensive index is greater than the preset threshold, the time period corresponding to the time window is determined to be an effective load event.

3. The built-in self-sensing intelligent cable health status monitoring method according to claim 2, characterized in that, Determine the load transfer relationship between multiple sensor measuring points, including: The sensor measuring points are regarded as nodes in a mechanical network. Based on the strain increment of one sensor measuring point and the strain increment of another sensor measuring point, the load transfer coefficient from the previous sensor measuring point to the next sensor measuring point is determined. The load transfer relationship matrix is ​​formed by the load transfer coefficients between multiple sensor measuring points.

4. The built-in self-sensing intelligent cable health status monitoring method according to claim 1, characterized in that, For the aforementioned effective load event, by analyzing the propagation characteristics of stress waves within the cable and the strain response characteristics, the axial and circumferential locations of cable damage are determined, including: Under the aforementioned effective load event, the actual arrival time of the stress wave from the excitation point to each sensor measurement point is obtained; The theoretical arrival time is determined based on the spatial location of each sensor measurement point. Based on the difference between the actual arrival time and the theoretical arrival time of each sensor point, sensor points with abnormal arrival time delays are selected to determine the axial and circumferential locations of cable damage.

5. The built-in self-sensing intelligent cable health status monitoring method according to claim 4, characterized in that, Determine the axial and circumferential locations of cable damage, including: Extract the response curves of sensor measurement points selected as having anomalies in arrival time delay during the strain rise phase; Analyze the variation characteristics of the response curve to determine the path complexity index that characterizes the complexity of the stress transmission path; Based on the sensor measurement points of the arrival time delay anomaly and the path complexity index, the axial and circumferential locations of the cable damage are determined.

6. The built-in self-sensing intelligent cable health status monitoring method according to claim 1, characterized in that, Based on the offset direction and distance of the monitoring data points relative to the cluster of healthy data points, the radial layer where the damage occurred is determined, including: If the first proportion value of the monitored data point is greater than the first proportion value of the health data point cluster center, and the difference between the second proportion value and the second proportion value of the health data point cluster center is within a preset range, then the damage is determined to occur in the inner layer. If the second proportional value of the monitored data point is less than the second proportional value of the health data point cluster center, and the difference between the first proportional value and the first proportional value of the health data point cluster center is within a preset range, then the damage is determined to occur in the outer layer. If the first proportion value of the monitored data point is greater than the first proportion value of the health data point cluster center, and the second proportion value of the monitored data point is less than the second proportion value of the health data point cluster center, then the damage is determined to occur in the middle layer.

7. The built-in self-sensing intelligent cable health status monitoring method according to claim 1, characterized in that, Identify health data point clusters, including: Under the healthy state of the cable, the first and second proportional values ​​corresponding to multiple complete load events are collected, and a data point distribution is formed in the two-dimensional phase space. The distribution of the data points is analyzed using a preset density estimation algorithm to determine the core region and boundary of the healthy data point cluster.

8. The built-in self-sensing intelligent cable health status monitoring method according to claim 1, characterized in that, This system is applied to a built-in self-sensing intelligent cable health status monitoring system, comprising: a layered built-in sensor network module, a data demodulation and acquisition module, and a data processing and analysis module; The layered built-in sensor network module includes multiple fiber optic strain sensors arranged radially in layers and axially in segments inside the cable, used to collect multi-channel strain data of the cable under load. The data demodulation and acquisition module is connected to the layered built-in sensor network module and is used to demodulate and acquire strain data from the fiber optic strain sensor in real time. The data processing and analysis module, which is communicatively connected to the data demodulation and acquisition module, is configured to execute: Based on the multi-channel strain data, identify effective load events; For the aforementioned effective load event, the axial and circumferential locations of cable damage are determined by analyzing the propagation characteristics of stress waves inside the cable and the strain response characteristics. The radial location and severity of cable damage are determined by the proportional relationship between strain data of different radial layers located at the same axial position.

Images