Composite material flexible pipe damage detection device and method

Abstract

The invention relates to the technical field of composite material flexible pipe detection, in particular to a composite material flexible pipe damage detection device and method. The detection device comprises a computer, and a fiber bragg grating demodulator, a dynamic signal test analysis system and a power amplifier which are connected with the computer, the pipeline model comprises a to-be-detected composite material flexible pipe, the composite material flexible pipe is horizontally placed and sleeved with a plurality of hoops at intervals, the bottom of the hoop on the left side is connected to a vibration exciter through a supporting leg, and the bottom of the hoop on the right side is connected to a plurality of fixing devices through supporting legs. The fiber bragg grating demodulator, the dynamic signal testing and analyzing system, the vibration exciter and the strain sensor are integrated, the integrated detection device is constructed, the composite material flexible pipe is fixed to be of a cantilever beam structure, controllable excitation is applied, high-precision and real-time collection of the strain response of the pipeline under the dynamic load is achieved, and the detection accuracy is improved. The problem that a traditional method is poor in adaptability to a non-rigid structure is effectively solved, and signal transmission reliability is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of composite flexible tube testing technology, specifically to a composite flexible tube damage testing device and method. Background Technology

[0002] Composite material flexible pipes, due to their lightweight, high strength, and corrosion resistance, are widely used in engineering fields such as petroleum, natural gas, and water conservancy, serving as key equipment for transporting fluid media. However, during long-term service, they are susceptible to various types of damage due to complex operating conditions (such as dynamic loads and media erosion), environmental factors (such as temperature and humidity changes), and external impacts. These damages include reinforcing layer cracks, interlayer damage, surface scratches, and localized voids, and are often scattered and multiple damages coexist. If these damages are not identified in time, they can spread over time, leading to pipeline leaks, ruptures, and other accidents, causing economic losses and environmental risks.

[0003] Currently, damage detection methods for composite material pipes, such as the patent technology in Chinese Patent Publication No. CN110131486A, mainly include manual inspection, ultrasonic inspection, X-ray inspection, and early data-driven inspection technologies, but all have significant limitations. Manual inspection relies on experience, is inefficient, struggles to detect hidden damage, and is unsuitable for complex structures and large-scale inspections. Ultrasonic and X-ray inspections have poor adaptability to the non-rigid structures of flexible pipes, and their detection accuracy is affected by pipe morphology, damage type, and inspection angle, making it difficult to locate and differentiate damage types. Early data-driven inspection technologies largely rely on numerical simulation data, but numerical simulations cannot fully reproduce the complex factors in actual engineering, such as noise interference, installation errors, and material property dispersion, leading to unstable detection accuracy in actual applications and failing to meet the reliability requirements of damage detection in engineering sites. Therefore, it is necessary to propose a damage detection device and method for composite material flexible pipes. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a device and method for detecting damage to flexible tubes made of composite materials.

[0005] The technical solution adopted in this invention is as follows: A composite flexible pipe damage detection device includes a horizontally placed pipe model and a detection device located on one side of the pipe model for detecting and analyzing damage to the pipe model, wherein: The detection device includes a computer and a fiber Bragg grating demodulator, a dynamic signal testing and analysis system, and a power amplifier connected to the computer. The fiber Bragg grating demodulator is positioned with strain sensors pointing towards the damaged area of ​​the pipe model. The data detected by the strain sensors is transmitted to the fiber Bragg grating demodulator, which feeds the signal back to the computer. The computer analyzes the signal via the dynamic signal testing and analysis system and transmits it to the data acquisition unit for storage. The obtained data is first stored and then amplified by the power amplifier. The pipeline model includes a composite flexible pipe to be tested, which is placed horizontally and has several clamps spaced on it. The bottom of the clamp on the left is connected to a vibration exciter via a support leg, which is used to adjust the vibration of the composite flexible pipe. The bottom of the clamp on the right is connected to several fixing devices via a support leg, which are used to maintain the cantilever beam state of the composite flexible pipe. The vibration exciter continuously applies dynamic excitation to the composite flexible tube to simulate vibration conditions in actual working conditions; strain sensors are set at several sampling points along the composite flexible tube to capture the strain response data of the pipe under different vibration conditions; the data obtained from the sampling points are fed back to the computer, which locates the damage location and classifies the damage type.

[0006] This technical solution constructs an integrated detection device comprising a fiber optic grating demodulator, a dynamic signal testing and analysis system, a vibration exciter, and a strain sensor. It applies controllable excitation by fixing a composite flexible tube as a cantilever beam structure, achieving high-precision, real-time acquisition of the pipeline's strain response under dynamic loads. The use of fiber optic strain sensors, combined with a dual fixing method of specialized adhesive and paper tape, ensures stable adhesion of the sensors to the flexible tube surface and reliable signal transmission, effectively overcoming the poor adaptability of traditional ultrasonic and X-ray methods to non-rigid structures. Simultaneously, it obtains strain data including actual noise, material dispersion, and installation errors through real physical experiments, avoiding the insufficient generalization ability of models caused by pure numerical simulation data, significantly improving the engineering credibility and field applicability of damage detection results.

[0007] In addition, the composite material flexible tube damage detection device proposed according to the present invention may also have the following additional technical features: According to one embodiment of the present invention, the composite flexible tube includes an inner liner and a reinforcing layer. The inner liner is made of PE material, has an inner diameter of 40 mm and a wall thickness of 4 mm. The reinforcing layer is composed of glass fiber and embedded in the PE matrix, wrapped around the inner liner at an angle of ±55°, with a total thickness of 1.5 mm. The inner liner and the reinforcing layer are bonded together by heating.

[0008] In this technical solution, the inner lining of the composite flexible pipe is made of PE material, which has good flexibility and corrosion resistance. The reinforcing layer consists of glass fiber embedded in the PE matrix. The glass fiber has high strength and high modulus characteristics and is wrapped around the inner lining at ±55° angles, so that the pipe is subjected to uniform stress in all directions, enhancing the overall strength and deformation resistance of the pipe. The inner lining and the reinforcing layer are bonded together by heating, so that the two are tightly combined to form an integral structure, which takes advantage of both the flexibility of PE material and the strength of glass fiber.

[0009] According to one embodiment of the present invention, the computer has the following built-in modules: The damage localization module determines the location of damage by analyzing the spatial distribution characteristics of strain data. The damage classification module distinguishes different damage types, including cracks, scratches, and delamination, by using the temporal characteristics and probability output of strain data.

[0010] In this technical solution, the computer's built-in damage localization module is based on the spatial distribution pattern of strain data. When a pipeline is damaged, the strain distribution in the damaged area and surrounding area differs from that in the normal area. By performing pattern recognition and feature extraction on the distribution patterns of a large amount of strain data in the spatial coordinate system, neural network classification is used to determine the specific location of the damage. The damage classification module utilizes the temporal characteristics of strain data, i.e., the law of strain change over time after damage occurs, as well as probabilistic statistical methods. Different types of damage have different temporal manifestations in their impact on strain, and each type of damage has corresponding probability distribution characteristics. Through comparative analysis, the damage types can be distinguished.

[0011] According to one embodiment of the present invention, the computer is also connected to an external display, which is used to display the detection results after computer analysis and processing in real time, including damage location, damage type and corresponding strain response data.

[0012] In this technical solution, the display can show the test results in real time, allowing users to understand the pipeline damage status immediately and take timely countermeasures; the data is presented in the form of images and text, which is more intuitive and easier to understand than simple numbers, reducing the difficulty of viewing; it reduces the time users spend analyzing data and improves the efficiency of overall testing and maintenance decisions.

[0013] According to one embodiment of the present invention, the strain sensor is a fiber optic grating sensor, which forms a narrowband filter by the photosensitivity of the fiber material and the change of the axial refractive index of the fiber core. When vibration causes the pipe to deform, the strain and grating period of the fiber optic grating change, thereby causing a change in the characteristic wavelength. The probe end of the strain sensor is attached to the surface of the composite flexible tube, and the position between adjacent sampling points is the location of the damage to be detected.

[0014] In this technical solution, the axial refractive index of the optical fiber core undergoes periodic changes, forming a narrowband filter structure. When external vibrations deform the flexible composite tube, the strain sensor attached to its surface generates strain, causing a change in the fiber grating period. Since there is a specific relationship between the grating period and the characteristic wavelength, the periodic change triggers a corresponding change in the characteristic wavelength. By detecting this wavelength change, the strain of the pipeline can be accurately sensed, and the location of the damage can be determined based on the correlation between strain and damage.

[0015] According to one embodiment of the present invention, the center wavelength of the strain sensor forming a narrowband filter is expressed as:

[0016] In the formula: The effective refractive index of the optical fiber, For the grating period; Vibration causes deformation in the pipe model, leading to changes in fiber optic grating strain and period, which in turn alters the characteristic wavelength. Temperature also affects the corresponding variables, causing changes in wavelength. Represented as:

[0017] In the formula: It is the coefficient of thermal expansion. It is the thermo-optical coefficient. It is the change in temperature. It refers to strain, specifically the optical strain coefficient of fiber optic cables. Represented as:

[0018] In the formula: and It is the strain optical tensor component. For the ratio of slack; During the experiment, it was assumed that the temperature change was not significant, i.e. The strain calculation formula at this time is expressed as: .

[0019] In this technical solution, the effective refractive index of the optical fiber and the grating period jointly determine the center wavelength of the narrowband filter. When the pipeline is vibrated and deformed, the fiber grating attached to it will generate strain, which will change the grating period and thus the characteristic wavelength. Temperature affects the wavelength through thermal expansion and thermo-optic effect. Thermal expansion changes the length of the optical fiber, and the thermo-optic effect changes the refractive index of the optical fiber. However, when the temperature change is not significant, the temperature effect can be ignored. By measuring the change in wavelength, the strain can be calculated according to the formula of strain and wavelength, thus solving the pipeline deformation problem.

[0020] According to one embodiment of the present invention, one end of the power amplifier is connected to a computer via a circuit, and the other end is connected to a vibration exciter via a circuit; the power amplifier provides a random excitation to the vibration exciter, thereby driving the composite flexible tube to vibrate, and the strain of the composite flexible tube fluctuates stably within a certain value range without external excitation.

[0021] In this technical solution, the control signal output by the computer has relatively low power and cannot directly drive the vibration exciter to generate vibration of sufficient intensity. After receiving the computer signal, the power amplifier amplifies its parameters so that the output signal has sufficient power to drive the vibration exciter. The vibration exciter converts the electrical signal into mechanical vibration, which drives the composite flexible tube to vibrate.

[0022] To achieve the above objectives, the present invention also provides a method for detecting damage to composite flexible tubes.

[0023] A method for detecting damage to composite flexible tubes includes the following steps: S1. The composite flexible tube is fixed into a cantilever beam structure using fixing devices and clamps. The bottom of the clamp on the left is connected to the vibration exciter via a support leg, and the bottom of the clamp on the right is connected to several fixing devices via a support leg. Strain sensors are attached to preset measuring points. The position between adjacent sampling points is the location of the damage to be detected. The strain sensors are fiber optic grating sensors, and their detection ends are attached to the surface of the composite flexible tube. Connect the fiber optic grating demodulator, data acquisition unit, power amplifier, vibration exciter, and computer. S2. In the dynamic signal testing and analysis system, parameters are set, the vibration exciter is started, and a random excitation is given to the vibration exciter through the power amplifier to drive the composite flexible tube to vibrate, simulating the vibration conditions in actual working conditions; the strain sensor captures the strain response data of the pipeline under different vibration conditions, which is demodulated by the fiber optic demodulator and then transmitted to the computer for storage by the data acquisition unit; under no external excitation, the strain of the composite flexible tube fluctuates stably within a certain range, while under excitation, the strain data changes; S3. The collected strain dataset is normalized and randomized, and the processed data is input into the damage localization model and damage classification model in the computer respectively. The damage localization module analyzes the spatial distribution characteristics of the strain data, and the damage classification module outputs the results based on the temporal characteristics and probability of the strain data. The two models work together to obtain preliminary damage information. S4. Determine the location of damage based on the spatial response characteristics output by the damage localization model, and determine the damage type based on the maximum probability category output by the damage classification model, including cracks, scratches, and delamination; the monitor displays the detection results after computer analysis and processing in real time, including damage location, damage type, and corresponding strain response data.

[0024] This technical solution achieves automatic identification and accurate discrimination of multiple types of damage to composite flexible tubes by normalizing and randomizing the collected strain data and constructing damage localization and classification models. This solution can not only accurately determine the location of damage through spatial strain distribution characteristics, but also effectively distinguish damage types based on the probability results output by the classification model, solving the problem that traditional detection methods struggle to balance localization accuracy and classification capability. Combined with high sampling frequency and multi-sensor collaborative acquisition, it enhances the system's sensitivity to minor damage and concurrent damage, providing efficient and intelligent technical support for the health monitoring and preventative maintenance of composite flexible tubes. Specifically, by constructing a cantilever beam structure from a composite flexible pipe using fixing devices and clamps, fiber optic strain sensors are attached to preset measuring points to capture the pipe's strain response data with high precision. Various instrument and equipment lines are connected to ensure smooth data transmission. After setting parameters in the dynamic signal testing and analysis system, a vibration exciter is activated, providing random excitation via a power amplifier to simulate actual working conditions. This allows the strain sensors to acquire strain data under different vibrations, which is then demodulated, acquired, and stored in a computer. The acquired strain data is then normalized and randomized, input into a damage localization and classification model. The localization model analyzes spatial distribution characteristics, while the classification model outputs results based on temporal characteristics and probability. These two models work together to obtain preliminary damage information. Finally, the damage location and type are determined based on the model output, and the detection results are displayed in real time, achieving accurate detection and intuitive display of damage to the composite flexible pipe.

[0025] According to one embodiment of the present invention, in step S2, parameters are set in the dynamic signal test and analysis system, the sampling frequency is set to 1000Hz, the wavelength range and interval are set according to actual needs, the strain conversion formula is input, the data storage time is set to 20s, the zero point of the power amplifier is calibrated, and the excitation force amplitude range of the vibration exciter is set to ±200N.

[0026] This technical solution uses a sampling frequency of 1000Hz, acquiring 1000 data points per second, which captures rapidly changing signal details and avoids signal distortion. The wavelength range and interval are set according to actual needs. By inputting the strain conversion formula, the raw signal detected by the sensor is converted into strain data; the data storage time is set to 20s to record enough data for analysis; the power amplifier zero point is calibrated to ensure accurate output signal; the vibration exciter's excitation force amplitude range is set to ±200N to simulate vibration excitation of different intensities in actual working conditions.

[0027] According to an embodiment of the present invention, the collected strain dataset in step S3 is normalized and randomized, and the data is divided into training set, validation set and test set according to the ratio of 80%:10%:10%. The strain data collected under a single excitation has a dimension of [20000,8], which is used to expand the dataset for damage localization. The dataset used for damage classification has a dimension of [300000,8] after repeated collection and merging under multiple working conditions.

[0028] This technical solution divides the dataset proportionally: the training set is used for the model to learn data features and patterns, the validation set is used to adjust model hyperparameters and evaluate model performance, and the test set is used to finally evaluate the model's performance on unknown data. The different dimensions of the dataset—data from a single stimulus is used for localization expansion, and data collected and merged repeatedly under multiple operating conditions is used for classification—comprehensively cover different damage conditions and improve the model's ability to identify various types of damage.

[0029] Compared with the prior art, the present invention has the following advantages: (1) The detection device of the present invention integrates a fiber optic grating demodulator, a dynamic signal testing and analysis system, a vibration exciter and a strain sensor to construct an integrated detection device. The composite flexible tube is fixed as a cantilever beam structure to apply controllable excitation, so as to realize high-precision and real-time acquisition of the strain response of the pipeline under dynamic load. (2) The detection device of the present invention uses a fiber optic strain sensor, which is fixed by a combination of special glue and paper tape to ensure that the sensor is stably attached to the surface of the flexible tube, effectively overcoming the problem of poor adaptability of traditional methods to non-rigid structures and ensuring the reliability of signal transmission. (3) The detection method of the present invention collects strain data, and after normalization and randomization, it is input into the damage location and classification model. The location module analyzes the spatial distribution, and the classification module outputs based on the temporal characteristics and probability, which can accurately determine the damage location and type. (4) The detection method of the present invention combines high sampling frequency with multi-sensor collaborative acquisition, which enhances the system’s sensitivity to minor damage and multiple damage concurrent situations, and provides efficient and intelligent technical support for health monitoring and preventive maintenance of composite flexible tubes. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of the detection device of the present invention.

[0031] Figure 2 This is an electrical connection diagram of the detection device of the present invention.

[0032] Figure 3 This is a wavelength interface diagram of a single-channel fiber Bragg grating sensor.

[0033] Figure 4 This is a schematic diagram of strain response data measured by a fiber Bragg grating sensor.

[0034] Figure 5 This is a schematic diagram of the damage location and measuring point location of the present invention.

[0035] Figure 6 This is a strain data graph of the measurement points corresponding to the five damage conditions of this invention.

[0036] In the diagram: 1. Fiber Bragg grating demodulator; 2. Dynamic signal testing and analysis system; 3. Power amplifier; 4. Computer; 5. Vibration exciter; 6. Fixing device; 7. Composite material flexible tube; 8. Clamp; 9. Strain sensor; 10. Data acquisition device. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] Example 1 like Figure 1 and Figure 2 As shown, this embodiment provides a composite material flexible pipe damage detection device, including a horizontally placed pipe model and a detection device located on one side of the pipe model, used to detect and analyze damage to the pipe model, wherein: The detection device includes a computer 4 and a fiber Bragg grating demodulator 1, a dynamic signal testing and analysis system 2, and a power amplifier 3 connected to the computer 4. The fiber Bragg grating demodulator 1 is positioned towards the damaged area of ​​the pipe model via a strain sensor 9. The data detected by the strain sensor 9 is transmitted to the fiber Bragg grating demodulator 1, which feeds back the signal to the computer 4. The computer 4 analyzes the signal via the dynamic signal testing and analysis system 2 and transmits it to the data acquisition unit 10 for storage. The obtained data is first stored and then amplified by the power amplifier 3. The pipeline model includes a flexible composite pipe 7 to be tested, which is placed horizontally and fitted with several clamps 8 at intervals. The bottom of the left clamp 8 is connected to a vibration exciter 5 via a support leg, which is used to adjust the vibration of the flexible composite pipe 7. The bottom of the right clamp 8 is connected to several fixing devices 6 via a support leg, which are used to maintain the cantilever beam state of the flexible composite pipe 7. The vibration exciter 5 continuously applies dynamic excitation to the flexible composite pipe 7 to simulate the vibration conditions in actual working conditions. Strain sensors 9 are set with several sampling points along the flexible composite pipe 7 to capture the strain response data of the pipeline under different vibration conditions. The data obtained from the sampling points is fed back to the computer 4, which locates the damage location and classifies the damage type.

[0039] In this embodiment, the detection device forms a stable cantilever beam structure by fixing one end of the composite flexible tube 7 to the fixing device 6 and connecting the other end to the vibration exciter 5, which facilitates the application of controllable dynamic excitation and effectively simulates the vibration environment in actual service. A fiber optic strain sensor 9 is used and double-fixed to the sampling point with special adhesive and paper tape, significantly improving the adhesion and strain transmission accuracy of the strain sensor 9 on the surface of the composite flexible tube 7, ensuring long-term, stable, and reliable data acquisition. Furthermore, the fiber optic demodulator 1, in conjunction with the series-connected strain sensors 9, achieves efficient synchronous acquisition of multi-channel signals and transmits the data to the computer 4 through the data acquisition unit 10, simplifying system wiring and improving electromagnetic interference resistance. The power amplifier 3, controlled by the computer 4, drives the vibration exciter 5, achieving precise control of the excitation signal and enhancing experimental repeatability and data consistency. The computer 4 has a pre-set deep learning model for damage identification, which can intelligently analyze the acquired strain data, automatically and accurately locate the damage position and classify the damage type, greatly improving detection efficiency and intelligence, overcoming the shortcomings of traditional methods that rely on manual interpretation or have insufficient detection accuracy.

[0040] Example 2 The following is a specific case analysis of the detection device in Example 1.

[0041] The composite flexible pipe 7 analyzed in this experiment has a total length of 2m, including an inner lining and a reinforcing layer. The inner lining is made of PE material with an inner diameter of 40mm and a wall thickness of 4mm. The reinforcing layer is composed of glass fiber embedded in the PE matrix and wrapped around the inner lining at ±55° angles, with a total thickness of 1.5mm. The inner lining and reinforcing layer are bonded together by heating. The inner lining of the composite flexible pipe 7 is made of PE material, which has good flexibility and corrosion resistance. The reinforcing layer is composed of glass fiber embedded in the PE matrix. Glass fiber has high strength and high modulus characteristics. Wrapped around the inner lining at ±55° angles, it ensures that the pipe is subjected to uniform stress in all directions, enhancing the overall strength and deformation resistance of the pipe. The inner lining and reinforcing layer are bonded together by heating, forming a tight bond and an integral structure, which utilizes both the flexibility of PE material and the strength advantage of glass fiber.

[0042] One end of the power amplifier 3 is connected to the computer 4 via a circuit, and the other end is connected to the vibration exciter 5 via a circuit. The power amplifier 3 provides a random excitation to the vibration exciter 5, thereby causing the composite flexible tube 7 to vibrate. Without external excitation, the strain of the composite flexible tube 7 fluctuates stably within a certain numerical range. In this technical solution, the control signal output by the computer 4 has relatively low power and cannot directly drive the vibration exciter 5 to generate sufficiently strong vibrations. After receiving the signal from the computer 4, the power amplifier 3 amplifies its parameters to ensure the output signal has sufficient power to drive the vibration exciter 5. The vibration exciter 5 converts the electrical signal into mechanical vibration, causing the composite flexible tube 7 to vibrate. It should be noted that this experimental model does not include the outer protective layer. Considering that the reinforcing layer of the composite flexible tube 7 is highly susceptible to damage during service, the damage simulated in this experiment mainly targets the reinforcing layer.

[0043] The strain sensor 9 is a fiber optic grating sensor. It utilizes the photosensitivity of the fiber material and the change in the axial refractive index of the fiber core to form a narrowband filter. When vibration causes pipe deformation, the strain and grating period of the fiber optic grating change, leading to a change in the characteristic wavelength. The detection end of the strain sensor 9 is attached to the surface of the composite flexible tube 7, and the position between adjacent sampling points indicates the location of the damage to be detected. In this technical solution, the axial refractive index of the fiber core changes periodically, forming a narrowband filter structure. When external vibration causes the composite flexible tube 7 to deform, the strain sensor 9 attached to its surface generates strain, causing a change in the fiber optic grating period. Since there is a specific relationship between the grating period and the characteristic wavelength, the period change triggers a corresponding change in the characteristic wavelength. By detecting this wavelength change, the strain of the pipe can be accurately sensed, and the damage location can be determined based on the correlation between strain and damage.

[0044] Among them, the strain sensor 9, which uses a fiber Bragg grating sensor, is the key component for acquiring strain data at different measuring points. The working principle of the fiber Bragg grating sensor is as follows: A fiber Bragg grating (FBG) forms a narrowband filter based on the photosensitivity of the fiber material and the change in the axial refractive index of the fiber core. During the reception of broadband light by the fiber Bragg grating, if a wavelength reaches the Bragg condition of the fiber Bragg grating, reflection will occur, while other wavelengths continue to propagate.

[0045] According to one embodiment of the present invention, the center wavelength of the narrowband filter formed by the strain sensor 9 is expressed as:

[0046] In the formula: The effective refractive index of the optical fiber, For the grating period; Vibration causes deformation in the pipe model, leading to changes in fiber optic grating strain and period, which in turn alters the characteristic wavelength. Temperature also affects the corresponding variables, causing changes in wavelength. Represented as:

[0047] In the formula: It is the coefficient of thermal expansion. It is the thermo-optical coefficient. It is the change in temperature. It refers to strain, specifically the optical strain coefficient of fiber optic cables. Represented as:

[0048] In the formula: and It is the strain optical tensor component. For the ratio of slack; During the experiment, it was assumed that the temperature change was not significant, i.e. The strain calculation formula at this time is expressed as: .

[0049] In this technical solution, the effective refractive index of the optical fiber and the grating period jointly determine the center wavelength of the narrowband filter. When the pipeline is vibrated and deformed, the fiber grating attached to it will generate strain, which will change the grating period and thus the characteristic wavelength. Temperature affects the wavelength through thermal expansion and thermo-optic effect. Thermal expansion changes the length of the optical fiber, and the thermo-optic effect changes the refractive index of the optical fiber. However, when the temperature change is not significant, the temperature effect can be ignored. By measuring the change in wavelength, the strain can be calculated according to the formula of strain and wavelength, thus solving the pipeline deformation problem.

[0050] like Figure 1As shown, one end of the composite flexible tube 7 is supported by two fixing devices 6, and the other end is bound to the composite flexible tube 7 by a fixed vibration exciter 5. Simultaneously, this experiment uses six fiber optic grating sensors to measure strain. Each sensor corresponds to a different wavelength and is connected in series, occupying only one signal channel. The sensors are attached to the reinforcement layer at equal intervals along the tube. Based on this, the fiber optic demodulator 1, dynamic signal testing and analysis system 2, and computer 4 operating software are turned on. The operating software interface is shown below. Figure 3 As shown, a certain sensor wavelength on the display channel is then zeroed and the measurement frequency is adjusted to 100Hz. After setting the working voltage, a signal is given to the dynamic signal test and analysis system 2, and a stopwatch is started. A random excitation is given to the exciter through the power amplifier 3, which in turn drives the pipe to vibrate. After a period of time (about 60 seconds), the timing is stopped, the data is saved, and the software is closed.

[0051] Experimental strain response data such as Figure 4 As shown, corresponding to the strain data from the four sensors, the pipe strain fluctuates stably within a certain range without external excitation. Figure 5 As shown, this study simulates damage by cutting cracks at different locations in the reinforcing layer of the composite flexible tube 7; this experimental study considers Figure 1 The analysis covers four different crack lengths and depths, resulting in a total of five damage scenarios: one intact scenario, one single-damage scenario, and three multi-damage scenarios. For details, see [link to relevant documentation]. Figure 3 Finally, for the five damage scenarios, 10-second strain response data from six measuring points were selected from the experimental data for analysis; the strain response data at measuring point 1 corresponding to different damage scenarios are shown below. Figure 6 Table 1 shows the four levels of crack damage, and Table 2 shows the specific damage conditions. Since the simulated crack damage is relatively small compared to the entire pipeline, it is difficult to directly distinguish the strain response data corresponding to different damage conditions.

[0052] Table 1 shows the damage levels of four types of cracks.

[0053] Table 2 shows the specific damage details.

[0054] In summary, the composite flexible pipe damage detection device and method, by rigidly fixing one end of the composite flexible pipe 7 with a fixing device 6 and connecting the other end to a vibration exciter 5, forms a stable cantilever beam structure, which can effectively apply controllable dynamic random excitation, realistically simulate the vibration environment of the pipeline during service, and improve the engineering representativeness of the experimental conditions; the use of fiber optic strain sensors 9, combined with a dual fixing method of sensor-specific adhesive and paper tape, to be pasted onto the preset measuring points of the reinforcement layer ensures effective strain transmission between the sensor and the pipe wall, significantly improving the accuracy and long-term stability of strain data acquisition; and the eight fiber optic strain sensors 9 are connected to the fiber optic demodulator 1 in a wavelength division multiplexing series manner to achieve single-channel multi-point synchronous acquisition, greatly simplifying wiring complexity, reducing system interference, and improving signal integrity.

[0055] Furthermore, by setting up computer 4 to send commands through dynamic signal testing and analysis system 2, which drive vibration exciter 5 via power amplifier 3, precise control and repeated loading of the excitation signal are achieved, ensuring the consistency and comparability of data under different damage conditions. Computer 4 has a pre-set deep learning model for damage identification, which can automatically extract features from the collected strain time-series data, determine the spatial location of damage through a damage localization model, and identify damage types (such as cracks, multiple damages, etc.) through a damage classification model, achieving a high degree of automation and intelligence in the detection process. By cutting cracks with defined lengths and depths at different locations in the reinforcement layer (as shown in Tables 1 and 2), an experimental dataset containing complete states and various single / multiple damage conditions is constructed, combined with strain response data at a sampling frequency of 100Hz (such as...). Figure 4 , Figure 6 As shown in the figure, the proposed method is systematically verified to demonstrate its effectiveness and robustness in identifying minute damage.

[0056] Example 3 Based on Example 1, the present invention also provides a method for detecting damage to composite flexible tubes, comprising the following steps: S1. The composite flexible tube 7 is fixed into a cantilever beam structure by fixing device 6 and clamp 8. The bottom of the clamp 8 on the left is connected to the vibration exciter 5 through a support leg, and the bottom of the clamp 8 on the right is connected to several fixing devices 6 through a support leg. The strain sensor 9 is attached to the preset measuring point. The position between adjacent sampling points is the damage location to be detected. The strain sensor 9 is a fiber optic grating sensor, and its detection end is attached to the surface of the composite flexible tube 7. The lines between the fiber optic grating demodulator 1, data acquisition unit 10, power amplifier 3, vibration exciter 5 and computer 4 are connected. S2. Parameter settings are made in the dynamic signal test and analysis system 2, the vibration exciter 5 is started, and a random excitation is given to the vibration exciter 5 through the power amplifier 3, which drives the composite flexible tube 7 to vibrate, simulating the vibration conditions in actual working conditions; the strain sensor 9 captures the strain response data of the pipeline under different vibration conditions, and after demodulation by the fiber optic demodulator 1, it is transmitted to the computer 4 for storage by the data acquisition unit 10; under no external excitation, the strain of the composite flexible tube 7 fluctuates stably within a certain range, while under excitation, the strain data changes; S3. The collected strain dataset is normalized and randomized, and the processed data is input into the damage localization model and damage classification model in computer 4 respectively. The damage localization module analyzes the spatial distribution characteristics of the strain data, and the damage classification module outputs the results through the temporal characteristics and probability of the strain data. The two models work together to obtain preliminary damage information. S4. Determine the location of the damage based on the spatial response characteristics output by the damage location model, and determine the damage type based on the maximum probability category output by the damage classification model, including cracks, scratches, and delamination; the monitor displays the detection results after analysis and processing by the computer in real time, including the damage location, damage type, and corresponding strain response data.

[0057] This technical solution achieves automatic identification and accurate discrimination of multiple types of damage in the composite flexible tube 7 by normalizing and randomizing the collected strain data and constructing damage localization and classification models. This solution can not only accurately determine the location of damage through spatial strain distribution characteristics, but also effectively distinguish damage types based on the probability results output by the classification model, solving the problem that traditional detection methods struggle to balance localization accuracy and classification capability. Combined with high sampling frequency and multi-sensor collaborative acquisition, it enhances the system's sensitivity to minor damage and concurrent damage, providing efficient and intelligent technical support for the health monitoring and preventative maintenance of the composite flexible tube 7. Specifically, by constructing a cantilever beam structure from the composite flexible pipe 7 using the fixing device 6 and clamps 8, fiber optic strain sensors 9 are attached to preset measuring points to capture the pipe's strain response data with high precision. Various instrument and equipment lines are connected to ensure smooth data transmission. After setting parameters in the dynamic signal testing and analysis system 2, the vibration exciter 5 is activated, and random excitation is provided through the power amplifier 3 to simulate actual working condition vibration, enabling the strain sensor 9 to acquire strain data under different vibrations. After demodulation and acquisition, the data is stored in the computer 4. Next, the acquired strain data is normalized and randomized, and input into the damage location and classification model. The location model analyzes spatial distribution characteristics, and the classification model outputs results based on temporal characteristics and probability. The two work together to obtain preliminary damage information. Finally, the damage location and type are determined based on the model output, and the detection results are displayed in real time on the monitor, achieving accurate detection and intuitive display of damage to the composite flexible pipe 7.

[0058] The deep learning model of this invention: A deep learning model for damage identification is preset in computer 4, which is the core intelligent part of this invention; The closed-loop detection process of this invention: Provides a complete detection method process, from equipment installation, parameter setting, data acquisition, data preprocessing, to finally using the deep learning model to achieve automated damage localization and classification, realizing a high degree of automation and intelligence in the detection process.

[0059] The computer 4 has the following built-in modules, which work together: The damage localization module determines the location of damage by analyzing the spatial distribution characteristics of strain data. The damage classification module distinguishes different damage types, including cracks, scratches, and delamination, by using the temporal characteristics and probability output of strain data.

[0060] In this technical solution, the damage location module built into the computer 4 is based on the spatial distribution pattern of strain data. When there is damage to the pipeline, the strain distribution of the damaged area and surrounding area is different from that of the normal area. By performing pattern recognition and feature extraction on the distribution pattern of a large amount of strain data in the spatial coordinate system, the specific location of the damage is determined by using neural network classification. The damage classification module utilizes the temporal characteristics of strain data, that is, the law of strain change over time after damage occurs, as well as probability statistics methods. Different types of damage have different temporal manifestations in their impact on strain, and each type of damage has corresponding probability distribution characteristics. Through comparative analysis, the damage types can be distinguished.

[0061] According to one embodiment of the present invention, the computer 4 is also connected to an external display, which is used to display the detection results after analysis and processing by the computer 4 in real time, including the damage location, damage type and corresponding strain response data.

[0062] In this technical solution, the display can show the test results in real time, allowing users to understand the pipeline damage status immediately and take timely countermeasures; the data is presented in the form of images and text, which is more intuitive and easier to understand than simple numbers, reducing the difficulty of viewing; it reduces the time users spend analyzing data and improves the efficiency of overall testing and maintenance decisions.

[0063] According to an embodiment of the present invention, in step S2, parameters are set in the dynamic signal test and analysis system 2, the sampling frequency is set to 1000Hz, the wavelength range and interval are set according to actual needs, the strain conversion formula is input, and the data storage time is set to 20s; the zero point of the power amplifier 3 is calibrated, and the excitation force amplitude range of the vibration exciter 5 is set to ±200N.

[0064] The sampling frequency of this technical solution is set to 1000Hz, acquiring 1000 data points per second, which can capture rapidly changing signal details and avoid signal distortion. The wavelength range and interval are set according to actual needs. By inputting the strain conversion formula, the raw signal detected by the sensor is converted into strain data; the data storage time is set to 20s to record enough data for analysis; the zero point of the power amplifier 3 is calibrated to ensure the accuracy of the output signal; the excitation force amplitude range of the vibration exciter 5 is set to ±200N to simulate vibration excitation of different intensities in actual working conditions.

[0065] According to an embodiment of the present invention, the collected strain dataset in step S3 is normalized and randomized, and the data is divided into training set, validation set and test set according to the ratio of 80%:10%:10%. The strain data collected under a single excitation has a dimension of [20000,8], which is used to expand the dataset for damage localization. The dataset used for damage classification has a dimension of [300000,8] after repeated collection and merging under multiple working conditions.

[0066] This technical solution divides the dataset proportionally: the training set is used for the model to learn data features and patterns, the validation set is used to adjust model hyperparameters and evaluate model performance, and the test set is used to finally evaluate the model's performance on unknown data. The different dimensions of the dataset—data from a single stimulus is used for localization expansion, and data collected and merged repeatedly under multiple operating conditions is used for classification—comprehensively cover different damage conditions and improve the model's ability to identify various types of damage.

[0067] In summary, the key to this invention lies in combining a structurally stable physical excitation platform, high-precision fiber optic sensing technology, and deep learning intelligent algorithms trained on real data, thereby solving problems such as low detection accuracy, difficulty in simultaneously addressing localization and classification, and insufficient model generalization ability in existing technologies.

[0068] Although the present invention has been described in detail with reference to the accompanying drawings and preferred embodiments, the invention is not limited thereto. Various equivalent modifications or substitutions can be made to the embodiments of the invention by those skilled in the art without departing from the spirit and essence of the invention, and such modifications or substitutions should all be within the scope of the invention. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the invention should also be covered within the protection scope of the invention. Therefore, the protection scope of the invention should be determined by the scope of the claims.

Claims

1. A device for detecting damage to flexible composite tubes, characterized in that, It includes a horizontally placed pipe model and a detection device located on one side of the pipe model, used to detect and analyze damage to the pipe model, wherein: The detection device includes a computer (4) and a fiber Bragg grating demodulator (1), a dynamic signal testing and analysis system (2), and a power amplifier (3) connected to the computer (4). The fiber Bragg grating demodulator (1) is set towards the damaged area of ​​the pipe model through a strain sensor (9). The data detected by the strain sensor (9) is transmitted to the fiber Bragg grating demodulator (1), and the fiber Bragg grating demodulator (1) feeds the signal back to the computer (4). The computer (4) analyzes the signal through the dynamic signal testing and analysis system (2) and transmits it to the data acquisition unit (10) for storage. The obtained data is first stored and then amplified by the power amplifier (3). The pipeline model includes a composite flexible pipe (7) to be tested, which is placed horizontally and has several clamps (8) spaced on it. The bottom of the clamp (8) on the left is connected to a vibration exciter (5) via a support leg. The vibration exciter (5) is used to adjust the vibration of the composite flexible pipe (7). The bottom of the clamp (8) on the right is connected to several fixing devices (6) via a support leg. The fixing devices (6) are used to maintain the cantilever beam state of the composite flexible pipe (7). Then, the vibration exciter (5) continuously applies dynamic excitation to the composite flexible tube (7) to simulate the vibration conditions in actual working conditions; the strain sensor (9) sets several sampling points along the composite flexible tube (7) to capture the strain response data of the pipe under different vibration conditions; the data obtained from the sampling points is fed back to the computer (4), and the computer (4) locates the damage location and classifies the damage type.

2. The composite material flexible tube damage detection device as described in claim 1, characterized in that, The composite flexible tube (7) includes an inner liner and a reinforcing layer. The inner liner is made of PE material with an inner diameter of 40 mm and a wall thickness of 4 mm. The reinforcing layer is composed of glass fiber and embedded in the PE matrix. It is wrapped around the inner liner at an angle of ±55° and has a total thickness of 1.5 mm. The inner liner and the reinforcing layer are bonded together by heating.

3. The composite material flexible tube damage detection device as described in claim 1, characterized in that, The computer (4) has the following built-in modules: The damage localization module determines the location of damage by analyzing the spatial distribution characteristics of strain data. The damage classification module distinguishes different damage types, including cracks, scratches, and delamination, by using the temporal characteristics and probability output of strain data.

4. The composite material flexible tube damage detection device as described in claim 1 or 3, characterized in that, The computer (4) is also connected to an external display, which is used to display the detection results after analysis and processing by the computer (4) in real time, including the damage location, damage type and corresponding strain response data.

5. The composite material flexible tube damage detection device as described in claim 1, characterized in that, The strain sensor (9) is a fiber optic grating sensor. It forms a narrowband filter by the photosensitivity of the fiber material and the change of the axial refractive index of the fiber core. When the vibration causes the pipe to deform, the strain and grating period of the fiber optic grating change, which in turn causes the characteristic wavelength to change. The probe end of the strain sensor (9) is attached to the surface of the composite flexible tube (7), and the position between adjacent sampling points is the location of the damage to be detected.

6. The composite material flexible tube damage detection device as described in claim 5, characterized in that, The center wavelength of the narrowband filter formed by the strain sensor (9) is expressed as: In the formula: The effective refractive index of the optical fiber, For the grating period; Vibration causes deformation in the pipe model, which leads to changes in the strain and period of the fiber grating, and thus changes the characteristic wavelength. Temperature also affects the corresponding variable, the amount of change in wavelength. Represented as: In the formula: It is the coefficient of thermal expansion. It is the thermo-optical coefficient. It is the change in temperature. It refers to strain, specifically the optical strain coefficient of fiber optic cables. Represented as: In the formula: and It is the strain optical tensor component. For the ratio of slack; During the experiment, it was assumed that the temperature change was not significant, i.e. The strain calculation formula at this time is expressed as: 。 7. The composite material flexible tube damage detection device as described in claim 1, characterized in that, One end of the power amplifier (3) is connected to the computer (4) via a line, and the other end is connected to the vibration exciter (5) via a line. The power amplifier (3) provides a random excitation to the vibration exciter (5), thereby driving the composite flexible tube (7) to vibrate. Under the absence of external excitation, the strain of the composite flexible tube (7) fluctuates stably within a certain range.

8. A method for detecting damage to composite flexible tubes, using the composite flexible tube damage detection device as described in any one of claims 1-7, characterized in that, Includes the following steps: S1. Fix the composite flexible tube (7) into a cantilever beam structure using fixing devices (6) and clamps (8). The bottom of the clamp (8) on the left is connected to the vibration exciter (5) via a support leg, and the bottom of the clamp (8) on the right is connected to several fixing devices (6) via a support leg. The strain sensor (9) is attached to the preset measurement point. The position between adjacent sampling points is the location of the damage to be detected. The strain sensor (9) is a fiber optic grating sensor, and its detection end is attached to the surface of the composite flexible tube (7). Connect the lines between the fiber optic grating demodulator (1), the data acquisition unit (10), the power amplifier (3), the vibration exciter (5), and the computer (4). S2. Set parameters in the dynamic signal test and analysis system (2), start the vibration exciter (5), and give the vibration exciter (5) a random excitation through the power amplifier (3) to drive the composite flexible tube (7) to vibrate, simulating the vibration conditions in actual working conditions; the strain sensor (9) captures the strain response data of the pipeline under different vibration conditions, and after demodulation by the fiber optic demodulator (1), it is transmitted to the computer (4) by the data acquisition unit (10) for storage; the strain of the composite flexible tube (7) under no external excitation is stable within a certain range, while the strain data changes under excitation. S3. Normalize and randomize the collected strain dataset, and input the processed data into the damage localization model and damage classification model in the computer (4); the damage localization module analyzes the spatial distribution characteristics of the strain data, and the damage classification module outputs the results through the temporal characteristics and probability of the strain data. The two models work together to obtain preliminary damage information. S4. Determine the location of damage based on the spatial response characteristics output by the damage location model, and determine the damage type based on the maximum probability category output by the damage classification model, including cracks, scratches, and delamination; the display shows the detection results after analysis and processing by the computer (4) in real time, including the damage location, damage type and corresponding strain response data.

9. The method for detecting damage to composite flexible tubes as described in claim 8, characterized in that, In the dynamic signal test and analysis system (2), parameters are set in S2, the sampling frequency is set to 1000Hz, the wavelength range and interval are set according to actual needs, the strain conversion formula is input, and the data storage time is set to 20s; the zero point of the power amplifier (3) is calibrated, and the excitation force amplitude range of the vibration exciter (5) is set to ±200N.

10. The method for detecting damage to composite flexible tubes as described in claim 8, characterized in that, The S3 process normalizes and randomizes the collected strain dataset, dividing it into training, validation, and test sets according to a ratio of 80%:10%:10%. The strain data collected under a single stimulus has a dimension of [20000,8], which is used to expand the dataset for damage localization. The dataset used for damage classification has a dimension of [300000,8] after repeated collection and merging under multiple working conditions.

Images