Steel structure fatigue crack monitoring device and method based on distributed optical fiber sound wave sensing technology

Abstract

The invention discloses a steel structure fatigue crack monitoring device and method based on a distributed optical fiber sound wave sensing technology, and belongs to the technical field of steel structure health state monitoring, the device structure comprises N optical fiber stress wave sensing units, a signal transmission optical fiber, a distributed optical fiber sound wave demodulator and an upper computer; the optical fiber stress wave sensing unit comprises a cylindrical elastic body, a spiral sensing optical fiber and a noise isolation shell; a spiral sensing optical fiber is wound on the outer side face of the cylindrical elastomer, and the noise isolation shell is used for wrapping the cylindrical elastomer and the spiral sensing optical fiber; the signal transmission optical fiber is connected with the optical fiber stress wave sensing units and the distributed optical fiber acoustic wave demodulator, and the spiral sensing optical fibers in the N optical fiber stress wave sensing units are connected in series; and the upper computer identifies crack initiation and expansion conditions based on the Rayleigh scattering light phase change signal. According to the device and the method, distributed, real-time and high-sensitivity monitoring on fatigue cracks of key parts such as main components and welding seams of a steel structure can be realized.

Description

Technical Field

[0001] This invention belongs to the field of steel structure health status monitoring technology, specifically relating to a steel structure fatigue crack monitoring device and method based on distributed fiber optic acoustic wave sensing technology. Background Technology

[0002] Steel structures are widely used in bridges, stadiums, large factories, marine engineering, and high-rise buildings. As engineering structures continue to develop towards larger spans and lighter weights, high-strength steel, due to its high strength, good toughness, and superior weldability, is widely used in various critical load-bearing components. However, under long-term loading, fatigue loads, and environmental factors (such as temperature changes, corrosion, and wind vibration), stress concentration points in structural components are still highly susceptible to fatigue cracks. For high-strength steel structures, although the material possesses high load-bearing capacity, its high yield strength ratio, high local brittleness sensitivity, and rapid fatigue crack propagation rate make it difficult to detect cracks in their early stages. These cracks often propagate rapidly within a short period, leading to a significant decrease in the load-bearing capacity of the components and even causing serious structural safety accidents. Therefore, real-time, long-term, and reliable online monitoring of fatigue cracks in steel structures, especially high-strength steel structures, is of significant engineering importance.

[0003] Existing fatigue crack monitoring methods mainly include ultrasonic testing, magnetic particle testing, acoustic emission monitoring, and resistance strain gauge testing. These methods all have certain limitations, such as long manual non-destructive testing cycles, reliance on operator experience, and inability to achieve continuous monitoring; the limited number and complex deployment of traditional acoustic emission sensors, making it difficult to achieve large-area coverage; and the poor durability and insufficient long-term stability of resistance strain gauges, which cannot effectively capture early-stage microcrack characteristic signals. Therefore, these methods are insufficient to meet engineering requirements in large steel structures, complex steel structures, concealed components, or long-term online monitoring scenarios.

[0004] Chinese patent document CN113899746A discloses a method for measuring the fatigue crack propagation morphology of steel structures based on DIC (Digital Image Conversion). This method includes: acquiring a sequence of digital images recording the changes in the fatigue crack propagation morphology of the steel structure; performing image processing on the digital image sequence to obtain a crack propagation displacement field with a topological structure; extracting the crack propagation morphology from the crack propagation displacement field to reconstruct the dynamic propagation process of the fatigue crack in the steel structure, thereby achieving high-precision and intelligent real-time monitoring of fatigue cracks. However, this method relies on stable observation conditions and is difficult to adapt to complex engineering site environments.

[0005] Chinese patent document CN116432475A discloses a multi-factor coupled collaborative early warning method for fatigue crack propagation in steel structures. This method includes: acquiring multi-physics monitoring data of the distribution points of hazard sources in steel structure engineering to obtain a monitoring time-series dataset; constructing an intuitionistic fuzzy matrix of the monitoring time-series dataset; using the grey relational coefficients between various physical field monitoring indicators to obtain the uncertainty of each indicator; assigning the obtained uncertainty as the basic probability value of each piece of evidence; preprocessing the evidence by weighted averaging to obtain a corrected basic probability value; obtaining the basic probability values ​​of the fatigue crack propagation process in different stages of the steel structure; and determining the fatigue crack propagation level of the distribution points of hazard sources in the steel structure engineering using basic probability assignment. This method has a complex algorithm and is overly dependent on the quality and real-time performance of the multi-source data.

[0006] With the development of fiber optic sensing technology, distributed fiber optic acoustic wave sensing technology has attracted attention in the field of structural health monitoring due to its advantages such as continuous distribution, high sensitivity, strong resistance to electromagnetic interference, and suitability for long-distance monitoring. Distributed fiber optic acoustic wave sensing technology can achieve continuous acoustic signal acquisition using ordinary single-mode optical fibers, enabling large-scale monitoring of weak stress waves, impacts, and abnormal vibrations generated by crack initiation and propagation, providing the possibility for early warning of fatigue cracks. However, due to the complex wave propagation characteristics of steel structures, noise interference, and the large scale of the measured area, existing distributed fiber optic acoustic wave sensing technology monitoring devices still have room for improvement in terms of crack detection sensitivity, positioning accuracy, and data analysis methods.

[0007] Therefore, there is an urgent need to develop a device and method that combines distributed fiber optic acoustic sensing technology and is suitable for monitoring fatigue cracks in steel structures, so as to realize early identification of microcracks in key parts, tracking of crack propagation process and real-time early warning, and improve the safety management capability of steel structures throughout their entire life cycle. Summary of the Invention

[0008] This invention aims to solve the technical problem of distributed, long-distance, and multi-point synchronous monitoring of fatigue cracks in steel structures in structural health monitoring. It provides a steel structure fatigue crack monitoring device based on distributed fiber optic acoustic wave sensing technology. The device of this invention can achieve distributed, real-time, and highly sensitive monitoring of fatigue cracks in key parts such as main components and welds of steel structures.

[0009] The specific technical solution adopted is as follows: A steel structure fatigue crack monitoring device based on distributed fiber optic acoustic wave sensing technology includes N fiber optic stress wave sensing units, signal transmission optical fiber, distributed fiber optic acoustic wave demodulator and host computer. The fiber optic stress wave sensing unit includes a cylindrical elastomer, a helical sensing fiber, and a noise isolation shell. The helical sensing fiber is wound around the outer surface of the cylindrical elastomer. One end of the cylindrical elastomer is attached to the surface of the steel structure under test to receive stress waves generated by the initiation and propagation of cracks in the steel structure. The helical sensing fiber and the cylindrical elastomer deform together under the action of stress waves for stress wave sensing. The noise isolation shell is used to cover the cylindrical elastomer and the helical sensing fiber to isolate the influence of external noise. The signal transmission fiber connects the fiber stress wave sensing unit and the distributed fiber acoustic demodulator, and connects the spiral sensing fibers in the N fiber stress wave sensing units in series. The distributed fiber optic acoustic demodulator is used to acquire the Rayleigh scattering phase change signal of the spiral sensing fiber in each fiber optic stress wave sensing unit during the test time period, and transmit it to the host computer. The host computer uses the Rayleigh scattering phase change signal to identify the crack initiation and propagation status in the monitoring area of ​​each fiber optic stress wave sensing unit during the test time period.

[0010] Furthermore, the cylindrical elastomer is made of a highly elastic, low-damping material, selected from polyurethane, polyamide, or composite polymer materials. The corresponding cylindrical elastomer can amplify the response to minute deformations caused by stress waves and uniformly apply these deformations to the helical sensing fiber, thereby improving signal detection sensitivity.

[0011] Furthermore, the noise isolation shell adopts a three-layer composite material structure, with an outer layer being a sound wave reflecting layer, a middle layer being a damping sound-absorbing material, and an inner layer being a thermosetting resin layer. The outer layer material is selected from any one of aluminum alloy, stainless steel, or glass fiber reinforced composite materials, and the middle layer material is selected from any one of polyurethane damping material, butyl rubber damping material, or polymer composite damping material, which is used to isolate environmental noise and mechanical vibration interference.

[0012] Furthermore, the noise isolation shell is provided with a reserved hole for the transmission optical fiber to pass through (the transmission optical fiber is fixed and sealed after passing through the reserved hole).

[0013] Furthermore, the host computer calculates the stress wave amplitude and characteristic parameters of each fiber stress wave sensing unit within the measured time period based on the Rayleigh scattering light phase signal, and identifies whether crack initiation or propagation occurs in the monitoring area.

[0014] This invention also provides a method for monitoring fatigue cracks in steel structures based on distributed fiber optic acoustic sensing technology. The method, utilizing the aforementioned monitoring device for monitoring fatigue cracks in steel structures based on distributed fiber optic acoustic sensing technology, specifically includes the following steps: N cascaded fiber optic stress wave sensing units are respectively attached to the surface of the steel structure under test. A distributed fiber optic acoustic demodulator is used to acquire the Rayleigh scattering phase change signal of the spiral sensing fiber in each fiber optic stress wave sensing unit during the test time period. Based on the Rayleigh scattering phase change signal, the crack initiation and propagation in the monitoring area are identified.

[0015] Furthermore, fiber optic stress wave sensing units can be distributed at arbitrary intervals in key areas of the steel structure under test, enabling full-domain crack monitoring and spatial positioning analysis of the steel structure through a distributed fiber optic acoustic demodulator.

[0016] Furthermore, the phase change of Rayleigh scattered light The following formula is used to calculate:

[0017] in, L g The length of the spiral sensing fiber. p e The effective photoelastic coefficient of the spiral sensing fiber. λ The wavelength of light. R Let Δ be the radius of the cylindrical elastic body. R This represents the change in radius of the cylindrical elastic body. α The angle between the spiral sensing fiber and the cylindrical elastomer along the axis.

[0018] The stress waves generated by the initiation and propagation of cracks in steel structures can be detected by causing radial deformation of the cylindrical elastic body at the picometer level. The monitoring method of this invention has the characteristics of high sensitivity.

[0019] Specifically, a deep learning classification model is used to classify the phase change signal of Rayleigh scattering light in the spiral sensing fiber of each fiber stress wave sensing unit during the measured time period, and to identify the crack initiation and propagation in the monitoring area of ​​each fiber stress wave sensing unit during the measured time period.

[0020] Furthermore, the Rayleigh scattering phase change signal is preprocessed to obtain a time-series signal. The time-series signal is then transformed by time-frequency to obtain time-spectrum data. The time-spectrum data obtained under different time window lengths and frequency resolutions are combined to construct multi-scale time-frequency features. These multi-scale time-frequency features are used as input to a deep learning classification model. The output of the deep learning classification model is the crack state classification result, which includes the crack-free state, crack initiation state, and crack propagation state. The deep learning classification model has a convolutional neural network structure to enhance the separability of crack initiation and propagation events.

[0021] Furthermore, the host computer performs spatial clustering and temporal evolution analysis on the identified crack initiation and propagation results, constructs a fatigue crack health index for steel structures, and displays the crack development status of the monitored area in a three-dimensional visualization format.

[0022] The device of this invention combines distributed fiber optic acoustic demodulation technology with fiber optic stress wave sensing unit. It uses a distributed fiber optic acoustic demodulator to acquire the phase change signal of Rayleigh scattered light from each monitoring unit and performs intelligent identification through a deep learning classification model to determine whether crack initiation or propagation has occurred, thereby achieving high-precision automated monitoring of fatigue damage in steel structures.

[0023] Compared with the prior art, the beneficial effects of the present invention are as follows: 1) By densely attaching helical sensing optical fibers to the outer surface of a cylindrical elastomer and utilizing distributed fiber optic acoustic demodulation technology to obtain the phase change of Rayleigh scattered light, the amplified sensing capability of picometer-level minute radial deformation is achieved. Compared with the limitations of traditional acoustic emission sensors, which are greatly affected by distance attenuation and difficult to deploy over a large area, this invention can achieve multi-point continuous distributed monitoring on a single optical fiber and capture high-frequency stress waves generated by crack initiation and propagation in real time, significantly improving the coverage and sensitivity of fatigue damage monitoring of steel structures.

[0024] 2) This invention utilizes the physical mechanism of phase change in a helical sensing fiber under stress wave action to construct a mapping model between phase change and micro-deformation of an elastic body. Furthermore, a deep learning classification algorithm is employed to automatically extract and identify time-frequency features. This method integrates the physical characteristics of stress waves with data-driven capabilities, effectively distinguishing between different types of events such as crack initiation, crack propagation, and environmental noise. It significantly improves the accuracy and anti-interference ability of crack identification, avoiding the problems of traditional threshold-based discrimination methods being susceptible to noise interference and having a high false positive rate.

[0025] 3) The fiber optic stress wave sensing unit of this invention is small in size and easy to install, and can be flexibly deployed at key locations such as welds, node areas, and stress concentration areas as needed. Its internal noise isolation shell structure effectively shields environmental noise and mechanical vibration, improving monitoring stability. Through a distributed fiber optic acoustic demodulator, each sensing unit can be connected in series on the same fiber to achieve large-scale multi-point synchronous monitoring, and supports visualization analysis of crack location, evolution trend, and health index by a host computer. The system features simple deployment, low maintenance, and controllable cost, demonstrating good engineering applicability and promising prospects for widespread application. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of a steel structure fatigue crack monitoring device based on distributed fiber optic acoustic sensing technology.

[0027] Figure 2This is a schematic diagram showing the deployment of the steel structure fatigue crack monitoring device of the present invention in the weld area of ​​a steel structure.

[0028] Figure 3 This refers to the phase change signal of Rayleigh scattered light detected by the method of this invention.

[0029] Figure 4 This is a diagram showing the identification results of the deep learning classification model in the method of this invention for the crack state of steel structures.

[0030] Figure descriptions: 1 Noise isolation shell, 2 Cylindrical elastomer, 3 Spiral sensing fiber, 4 Reserved hole, 5 Signal transmission fiber, 6 Distributed fiber optic acoustic demodulator, 7 Steel structure component, 8 Fiber optic stress wave sensing unit. Detailed Implementation

[0031] To make the objectives, features, and advantages of this invention more apparent and understandable, a detailed description is provided below through specific embodiments. Many specific details are set forth in the following description to provide a thorough understanding of the invention. However, the invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below. Technical features in various embodiments of the invention can be combined appropriately without mutual conflict.

[0032] Unless otherwise specified, the operating methods in the following examples are generally performed under conventional conditions or as recommended by the manufacturer. Contents not described in detail in this specification are prior art known to those skilled in the art. Unless otherwise specified, the experimental materials used in the examples below can be purchased from conventional biochemical reagent companies.

[0033] Example 1 like Figure 1 As shown in this embodiment, the steel structure fatigue crack monitoring device based on distributed fiber optic acoustic wave sensing technology includes N fiber optic stress wave sensing units, signal transmission fiber 5, distributed fiber optic acoustic wave demodulator 6, and host computer, where N is a positive integer ≥1.

[0034] Each fiber optic stress wave sensing unit includes: a noise isolation shell 1, a cylindrical elastomer 2, a helical sensing fiber 3, and a pre-drilled hole 4. The outer surface of the cylindrical elastomer 2 is wound with the helical sensing fiber 3. One end of the cylindrical elastomer 2 is used to attach to the surface of the steel structure to be tested to receive stress waves generated by the initiation and propagation of cracks in the steel structure. The helical sensing fiber 3 is attached to the surface of the cylindrical elastomer 2 in a densely wound form and deforms in tandem with the cylindrical elastomer 2 under the action of stress waves for stress wave sensing. The noise isolation shell 1 is used to cover the cylindrical elastomer 2 and the helical sensing fiber 3 to isolate the influence of external noise and prevent noise from reaching the cylindrical elastomer 2 and the helical sensing fiber 3. The noise isolation shell 1 is provided with a pre-drilled hole 4 for the transmission fiber 5 to pass through, and then the hole is sealed.

[0035] The cylindrical elastomer 2 is made of a highly elastic, low-damping material, which is selected from polyurethane, polyamide, or polymer composite materials. The noise isolation shell adopts a three-layer composite material structure, with an outer layer being a sound wave reflecting layer, a middle layer being a damping and sound-absorbing material, and an inner layer being a thermosetting resin layer. The outer layer material is selected from any one of aluminum alloy, stainless steel, or glass fiber reinforced composite materials, and the middle layer material is selected from any one of polyurethane damping material, butyl rubber damping material, or polymer composite damping material.

[0036] One end of the signal transmission optical fiber 5 is connected to the distributed optical fiber acoustic demodulator 6, and the other end of the signal transmission optical fiber 5 is connected to N optical fiber stress wave sensing units in sequence, and the spiral sensing optical fibers 3 in the N optical fiber stress wave sensing units are connected in series. The signal transmission optical fiber 5 and the reserved hole 4 are both fixed.

[0037] Because distributed fiber optic acoustic wave sensing technology has the capability to achieve distributed acoustic wave monitoring over tens to hundreds of kilometers along a single optical fiber, 500 to 5000 fiber optic stress wave sensing units can be connected in series on a single signal transmission fiber 5. 500 and 5000 correspond to fiber lengths of 10 kilometers and 100 kilometers, respectively. Adjacent fiber optic stress wave sensing units are spaced 10 meters apart, and the spiral sensing fiber 3 within each unit is 10 meters long. In practical applications, multiple monitoring units can be connected in series with a minimum spatial resolution to form a steel structure crack monitoring string, thus meeting the distributed monitoring requirements for large steel structure cracks.

[0038] The host computer is used to acquire the Rayleigh scattering phase change signal of the spiral sensing fiber 3 in each fiber stress wave sensing unit obtained by the distributed fiber acoustic demodulator 6 during the measured time period, and to calculate the stress wave amplitude and parameters in each fiber stress wave sensing unit during the measured time period based on the Rayleigh scattering phase change signal, and to identify the crack initiation and propagation in the monitoring area of ​​each fiber stress wave sensing unit during the measured time period.

[0039] The host computer is also used to execute a deep learning classification model, which is used to classify the phase change signal of Rayleigh scattering light of the spiral sensing fiber in each fiber stress wave sensing unit during the test time period, and to identify the crack initiation and propagation in the monitoring area of ​​each fiber stress wave sensing unit during the test time period.

[0040] When a steel structure fatigue crack monitoring device based on distributed fiber optic acoustic sensing technology is applied, when a crack initiation or propagation occurs in the steel structure, the stress wave generated by the crack initiation or propagation will act on the cylindrical elastic body 2 of the stress wave sensing unit, causing radial deformation of the cylindrical elastic body 2. This, in turn, causes the spiral sensing fiber 3 wound on the cylindrical elastic body 2 to undergo tensile deformation, resulting in a change in the phase of the Rayleigh scattered light transmitted in the spiral sensing fiber 3.

[0041] This embodiment utilizes fiber optic stress wave sensing units, encapsulated into distributed, real-time monitoring units suitable for the initiation or propagation of cracks in steel structures. Multiple fiber optic stress wave sensing units are connected in series to perform distributed, real-time monitoring of crack initiation and propagation in steel structures. When no crack initiation or propagation occurs in the monitoring area of ​​the fiber optic stress wave monitoring unit, the Rayleigh scattering phase signal in the helical sensing fiber 3 of the unit undergoes a slight change due to noise. When crack initiation or propagation occurs in the monitoring area, the stress wave generated by the crack initiation or propagation causes a significant change in the Rayleigh scattering phase signal in the helical sensing fiber 3. The signal is demodulated using a distributed fiber optic acoustic demodulator 6, and a deep learning classification model is employed to achieve distributed, real-time monitoring of crack initiation and propagation in steel structures.

[0042] A steel structure fatigue crack monitoring device based on distributed fiber optic acoustic sensing technology is used to monitor the initiation or propagation of fatigue cracks in steel structures. The specific monitoring method is as follows: N cascaded fiber optic stress wave sensing units are respectively attached to the surface of the steel structure under test (placed in areas of the steel structure prone to cracking). The phase change of the Rayleigh scattered light in the spiral sensing fiber 3 of one fiber optic stress wave sensing unit is calculated according to the following formula. :

[0043] in, L g The length of the spiral sensing fiber. p e The effective photoelastic coefficient of the spiral sensing fiber. λ The wavelength of light. R Let Δ be the radius of the cylindrical elastic body 2. R Let be the change in radius of the cylindrical elastic body 2.α The angle between the spiral sensing fiber 3 and the cylindrical elastomer 2 is axial.

[0044] By using a distributed fiber optic acoustic demodulator 6 to acquire the Rayleigh scattering phase change signal of the spiral sensing fiber in each fiber optic stress wave sensing unit during the measured time period, the phase-time curve of the Rayleigh scattering light measured by each fiber optic stress wave sensing unit can be obtained. The obtained phase-time curves of the Rayleigh scattering light measured by each fiber optic stress wave sensing unit are then input into a deep learning classification model to determine whether crack initiation or propagation exists in the monitoring area of ​​each fiber optic stress wave sensing unit.

[0045] Before inputting the Rayleigh scattering phase-time curves obtained from each fiber stress wave sensing unit into the deep learning classification model, the Rayleigh scattering phase-time curves are first preprocessed. The preprocessing process includes detrending the phase change signal, noise suppression, and amplitude normalization to reduce the impact of system noise and environmental interference on the signal analysis results and obtain a stable phase change time series signal.

[0046] Subsequently, the preprocessed phase-change time-series signal undergoes time-frequency transformation to convert the one-dimensional time-domain signal into a two-dimensional time-frequency representation. Specifically, by performing time-frequency analysis on the phase-change signal under different time window lengths and frequency resolutions, time-spectrum data reflecting the energy distribution characteristics of the signal at different times and frequency scales are obtained. The time-spectrum data obtained at multiple time and frequency scales are combined to construct a multi-scale time-frequency feature, which is used to comprehensively characterize the time-frequency characteristics of the acoustic signal during crack initiation and crack propagation.

[0047] Multi-scale time-frequency features are used as input to the deep learning classification model, and the output of the deep learning classification model is the crack state classification result, which is used to determine the crack state of the steel structure in the monitoring area of ​​the corresponding fiber optic stress wave sensing unit. The crack state classification result includes at least the crack-free state, crack initiation state, and crack propagation state.

[0048] The deep learning classification model adopts a convolutional neural network structure. By performing convolutional feature extraction and feature fusion on the multi-scale time-frequency features of the input, it realizes the automatic learning of the feature patterns of crack-related acoustic signals in the time-frequency domain, thereby enhancing the separability of different crack states in the feature space and improving the accuracy of crack initiation and crack propagation identification.

[0049] During the model training phase, fatigue loading tests on steel structures under laboratory conditions were conducted. Loading tests were performed on steel structural members in different crack states, and Rayleigh scattering phase change signals from each fiber optic stress wave sensing unit were simultaneously acquired under the corresponding working conditions. The acquired signals were preprocessed and time-frequency features were extracted using the methods described above. The sample data were labeled according to the actual crack state to construct a training sample set. A deep learning classification model was trained using this training sample set. The model parameters were trained and optimized through supervised learning to obtain a parameter-optimized deep learning classification model. This parameter-optimized deep learning classification model was used to determine the crack state based on the Rayleigh scattering phase-time curves acquired in actual engineering projects.

[0050] In practical applications, such as Figure 2 As shown, the series-connected fiber optic stress wave sensing units 8, calibrated in the laboratory, are fixed onto the steel structure component 7 that requires crack monitoring. The series-connected fiber optic stress wave sensing units 8 are then connected to a distributed fiber optic acoustic demodulator and a host computer. The signals from each fiber optic stress wave sensing unit 8 can be demodulated and analyzed using the distributed fiber optic acoustic demodulator and the host computer, thereby achieving the purpose of real-time monitoring of cracks in the steel structure.

[0051] Figure 3 When the method of this invention is applied to a steel structure, the Rayleigh scattering phase change signal of each fiber optic stress wave sensing unit within its monitoring area is monitored. Numbered 7-1 to 7-10, these signals correspond to... Figure 2 Monitoring areas 7-1 to 7-10. Figure 4 To be Figure 3 The identification results of the crack state in each region are obtained by inputting the Rayleigh scattering phase signals of each monitoring unit into a deep learning classification model. Figure 4 This allows us to obtain information on crack patterns in various areas of the steel structure.

[0052] The embodiments described above provide a detailed explanation of the technical solutions of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, additions, or similar substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A fatigue crack monitoring device for steel structures based on distributed fiber optic acoustic sensing technology, characterized in that, It includes N fiber optic stress wave sensing units, signal transmission optical fibers, a distributed fiber optic acoustic demodulator, and a host computer; The fiber optic stress wave sensing unit includes a cylindrical elastomer, a helical sensing fiber, and a noise isolation shell. The helical sensing fiber is wound around the outer surface of the cylindrical elastomer. One end of the cylindrical elastomer is attached to the surface of the steel structure under test to receive stress waves generated by the initiation and propagation of cracks in the steel structure. The helical sensing fiber and the cylindrical elastomer deform together under the action of stress waves for stress wave sensing. The noise isolation shell is used to cover the cylindrical elastomer and the helical sensing fiber to isolate the influence of external noise. The signal transmission fiber connects the fiber stress wave sensing unit and the distributed fiber acoustic demodulator, and connects the spiral sensing fibers in the N fiber stress wave sensing units in series. The distributed fiber optic acoustic demodulator is used to acquire the Rayleigh scattering phase change signal of the spiral sensing fiber in each fiber optic stress wave sensing unit during the test time period, and transmit it to the host computer. The host computer uses the Rayleigh scattering phase change signal to identify the crack initiation and propagation status in the monitoring area of ​​each fiber optic stress wave sensing unit during the test time period.

2. The steel structure fatigue crack monitoring device based on distributed fiber optic acoustic sensing technology according to claim 1, characterized in that, The cylindrical elastomer is made of a highly elastic, low-damping material, which is selected from polyurethane, polyamide, or polymer composite materials.

3. The steel structure fatigue crack monitoring device based on distributed fiber optic acoustic sensing technology according to claim 1, characterized in that, The noise isolation shell adopts a three-layer composite material structure. The outer layer is a sound wave reflecting layer, the middle layer is a damping sound absorbing material, and the inner layer is a thermosetting resin layer. The outer layer material is selected from any one of aluminum alloy, stainless steel or glass fiber reinforced composite material, and the middle layer material is selected from any one of polyurethane damping material, butyl rubber damping material or polymer composite damping material.

4. The steel structure fatigue crack monitoring device based on distributed fiber optic acoustic sensing technology according to claim 1, characterized in that, The noise isolation shell is provided with reserved holes for the installation of optical fibers for signal transmission.

5. A method for monitoring fatigue cracks in steel structures based on distributed fiber optic acoustic sensing technology, characterized in that, The steel structure fatigue crack monitoring device based on distributed fiber optic acoustic wave sensing technology as described in any one of claims 1-4 specifically includes the following steps: N cascaded fiber optic stress wave sensing units are attached to the surface of the steel structure under test. A distributed fiber optic acoustic demodulator is used to acquire the Rayleigh scattering phase change signal of the spiral sensing fiber in each fiber optic stress wave sensing unit during the test time period. Based on the Rayleigh scattering phase change data, the crack initiation and propagation in the monitoring area are identified.

6. The method for monitoring fatigue cracks in steel structures based on distributed fiber optic acoustic sensing technology according to claim 5, characterized in that, Rayleigh scattering phase change The following formula is used to calculate: ； in, L g The length of the spiral sensing fiber. p e The effective photoelastic coefficient of the spiral sensing fiber. λ The wavelength of light. R Let Δ be the radius of the cylindrical elastic body. R This represents the change in radius of the cylindrical elastic body. α The angle between the spiral sensing fiber and the cylindrical elastomer along the axis.

7. The method for monitoring fatigue cracks in steel structures based on distributed fiber optic acoustic sensing technology according to claim 5, characterized in that, A deep learning classification model is used to classify the phase change signal of Rayleigh scattering light in the spiral sensing fiber of each fiber stress wave sensing unit during the test time period, and to identify the crack initiation and propagation in the monitoring area of ​​each fiber stress wave sensing unit during the test time period.

8. The method for monitoring fatigue cracks in steel structures based on distributed fiber optic acoustic sensing technology according to claim 7, characterized in that, The input of the deep learning classification model is the multi-scale time-frequency feature constructed by performing time-frequency transformation on the phase change signal of Rayleigh scattering light, and the output is the crack state classification result. The structure of the deep learning classification model is a convolutional neural network structure.

9. The method for monitoring fatigue cracks in steel structures based on distributed fiber optic acoustic sensing technology according to claim 5, characterized in that, The host computer further performs spatial clustering and temporal evolution analysis on the identified crack initiation and propagation results, constructs a fatigue crack health index for steel structures, and displays the crack development status of the monitored area in a three-dimensional visualization form.

Images