Concrete filled steel tube interface stripping defect positioning and detecting system

Abstract

The invention provides a concrete-filled steel tube interface stripping defect positioning and detecting system, and relates to the technical field of structure monitoring. Piezoelectric ceramic sensors are installed on the outer wall of a steel tube of a concrete-filled steel tube member and the lower surface of an annular plate, an impedance analyzer is used for applying alternating voltage excitation, and the stripping defect of the concrete-filled steel tube interface is detected by means of the positive and negative piezoelectric effects of piezoelectric ceramic materials. And enabling the sensor to output an electromechanical coupling admittance frequency curve. Admittance data are processed and converted into picture data through a computer, the picture data are input into a convolutional neural network model, and accurate positioning of the concrete filled steel tube member interface stripping defect is achieved through feature extraction and classification of a convolutional layer, a pooling layer and a full connection layer. The system is particularly suitable for detecting local initial damage in a high-frequency range, has the characteristics of simplicity and convenience in operation, quick response and high visibility, and provides a quick and accurate solution for real-time detection of the concrete-filled steel tube member.

Description

Technical Field

[0001] The invention relates to the technical field of structure monitoring, and in particular to a steel tube concrete interface peeling defect positioning detection system. Background Art

[0002] Steel tube concrete components are widely used in structures such as super high-rise buildings and long-span bridges due to their excellent mechanical properties, such as high bearing capacity and good ductility. However, during the long-term use of such components, due to the shrinkage and creep of concrete and quality control problems during construction, the bonding interface between the inner wall of the steel tube and the core concrete and between the lower surface of the inner ring plate and the concrete often debonds. This debonding defect will significantly reduce the bearing capacity and stiffness of the component, thus posing a threat to the safety and functionality of the structure.

[0003] At present, the detection methods for the interface debonding defects of steel tube concrete components mainly include the tapping method, fiber optic sensing method, impact echo method and ultrasonic detection. The tapping method has many limitations in practical applications, such as poor detection effect on thick-walled steel tube concrete components, and it relies on the subjective experience of operators and lacks objective judgment criteria. Although fiber optic sensors can provide relatively accurate monitoring data, their survival rate in concrete is low, and the equipment cost is high and the operation is complicated. Although the impact echo method can achieve non-contact detection, its detection accuracy and reliability are limited when facing large-section or thick-walled steel tube concrete components, and it is not effective in detecting interface debonding defects on the lower surface of the inner ring plate. Traditional ultrasonic detection methods also have certain limitations in the detection of internal interface defects in steel tube concrete, and it is difficult to accurately identify and locate defects.

[0004] Therefore, the existing detection methods have many deficiencies in practical engineering applications and cannot meet the needs of efficient and accurate detection of interface debonding defects in steel tube concrete components. This has prompted researchers to seek new detection technologies to improve the detection efficiency and accuracy of interface debonding defects in steel tube concrete components and ensure the safety and reliability of the structure.

[0005] In view of this, this application is filed. Summary of the invention

[0006] The present invention provides a steel tube concrete interface peeling defect positioning detection system, which can at least partially improve the above-mentioned problem.

[0007] To achieve the above object, the present invention adopts the following technical solutions: A steel tube concrete interface peeling defect positioning detection system, comprising: a control component, a steel and concrete composite component, an inner ring plate, an external voltage sensor component, and an embedded voltage sensor component, wherein the inner ring plate is arranged inside the steel and concrete composite component, the external voltage sensor component is arranged on the outer wall surface of the steel and concrete composite component, the embedded voltage sensor component is arranged on the lower surface of the inner ring plate, and the data end of the control component is electrically connected to the data end of the external voltage sensor component and the data end of the embedded voltage sensor component; The control component is configured to implement the following steps by executing a computer program stored therein: Acquire detection data collected by the external voltage sensor component and the embedded voltage sensor component, and perform Euclidean calculation preprocessing on the detection data; The preprocessed detection data is input into the preset convolutional neural network model to obtain the defect location results.

[0008] In summary, the present invention proposes an innovative positioning detection system, which is specifically used to identify and locate the bonding interface debonding defects between the inner wall of the steel tube and the core concrete, and between the lower surface of the inner ring plate and the concrete in the steel tube concrete member. The system combines the electromechanical coupling admittance measurement technology of external and embedded piezoelectric sensors, as well as advanced convolutional neural network algorithms, to achieve efficient and accurate detection of interface debonding defects in steel tube concrete members.

[0009] Specifically, the system installs piezoelectric ceramic sensors at key positions of steel tube concrete components. These sensors can generate and receive signals on their own and obtain the sensor's admittance-frequency curve under the excitation of high-frequency alternating voltage. These curves are converted into image data and then input into the convolutional neural network model. The model performs deep learning and feature extraction on the data through a series of convolution, pooling and full connection operations, and finally determines the location of the defect through the Softmax classifier. The innovation of this system is that it can not only detect the interface debonding defects in steel tube concrete components, but also accurately locate these defects, which is crucial for maintaining the safety and functionality of the structure. Compared with traditional detection methods, this system shows higher accuracy and reliability in detecting thick-walled steel tube concrete components, especially in the high-frequency range. The sensitivity to local initial damage is higher. In addition, the system has a simple operation process, fast response speed, and extremely high visibility, making the detection results intuitive and easy to understand; at the same time, it performs well in identifying steel tube concrete interface debonding defects.

[0010] In simple terms, the steel tube concrete interface peeling defect positioning detection device pastes the piezoelectric material on the outer wall of the steel tube and the lower surface of the inner ring plate in the form of surface pasting or embedding, and forms an electromechanical coupling system with the steel tube concrete structure to be tested. The piezoelectric material's self-excitation and self-sensing function is used to couple the piezoelectric sensor's own electrical admittance with the main structure's mechanical admittance. Since the piezoelectric sensor's parameters and performance remain unchanged, the main structure's mechanical admittance is the only parameter that affects the electromechanical coupling admittance of the piezoelectric material. When peeling occurs at the interface between the inner wall of the steel tube and the core concrete, and peeling defects occur at the interface between the lower surface of the ring plate and the concrete in the steel tube concrete component, the mechanical impedance of the structure at that location changes, which is then reflected by the electromechanical coupling admittance of the piezoelectric sensor at that location. In the piezoelectric admittance detection technology, the electrical admittance of the piezoelectric sensor is coupled with the mechanical impedance of the main structure. By measuring the electrical admittance signal received by the piezoelectric sensor, the changes in structural characteristics such as structural stiffness or damping caused by interface debonding defects in the main structure of the steel tube concrete can be directly reflected, and then the location and identification of the interface debonding defects can be achieved by analyzing the electromechanical coupling admittance signal; by establishing a convolutional neural network to compare the admittance of the piezoelectric sensor installed in the steel tube concrete component, the change in admittance before and after the interface debonding can be realized, and then the interface bonding condition between the inner wall of the steel tube, the lower surface of the inner ring plate and the concrete in the steel tube concrete component can be qualitatively and quantitatively judged, thereby accurately determining the location of the interface debonding defect. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a structural schematic diagram of a steel tube concrete interface peeling defect positioning and detection system provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of the piezoelectric admittance detection structure of the steel tube concrete interface debonding defect positioning detection system provided by an embodiment of the present invention; Figure 3 It is a flow chart of a steel tube concrete interface peeling defect positioning and detection system provided by an embodiment of the present invention; Figure 4 is a schematic structural diagram of an embedded voltage sensor assembly provided by an embodiment of the present invention; Figure 5 is a schematic structural diagram of an externally mounted piezoelectric sensor assembly provided by an embodiment of the present invention; Figure 6 is a schematic diagram of the arrangement of an externally mounted piezoelectric sensor assembly provided in an embodiment of the present invention, wherein the range indicated by the dotted line is a peeling defect area; Figure 7 : is an admittance signal diagram provided by an embodiment of the present invention, which is detected by taking the interface debonding of a steel tube concrete component as an example, wherein the dotted line is the admittance signal measured by the piezoelectric sensor at the debonding defect position, and the solid line is the admittance signal at the healthy position; Figure 8is a schematic diagram of a convolutional neural network model provided by an embodiment of the present invention; Fig. 9 It is a schematic diagram of the detection results of the debonding defect positioning at the steel tube concrete interface using the convolutional neural network and piezoelectric admittance measurement provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0013] refer to Figures 1 to 3 As shown, the first embodiment of the present invention discloses a steel tube concrete interface peeling defect positioning detection system, which includes: a control component, a steel and concrete composite component 10, an inner ring plate 3, an external voltage sensor component 4, and an embedded voltage sensor component 5, wherein the inner ring plate 3 is arranged inside the steel and concrete composite component 10, the external voltage sensor component 4 is arranged on the outer wall surface of the steel and concrete composite component 10, and the embedded voltage sensor component 5 is arranged on the lower surface of the inner ring plate 3, and the data end of the control component is electrically connected to the data end of the external voltage sensor component 4 and the data end of the embedded voltage sensor component 5; Preferably, the steel-concrete composite component 10 includes a steel component module 1 and concrete 2 , and the concrete 2 is poured inside the steel component module 1 .

[0014] In this embodiment, a steel tube concrete interface peeling defect positioning detection device is constructed, and the system is intended to improve the detection efficiency and accuracy of the internal interface peeling defects of steel tube concrete components. Specifically, the steel and concrete composite component 10 is the basic structure of this system, which is composed of a steel component module 1 and concrete 2 poured inside it. This structural design makes full use of the constraint effect of the steel component module 1 on the concrete 2, and improves the bearing capacity and ductility of the component. The inner ring plate 3 is arranged inside the steel and concrete composite component 10, which plays a role in uniform force distribution. The external voltage sensor assembly 4 is installed on the outer wall surface of the steel and concrete composite component 10, and the embedded voltage sensor assembly 5 is installed on the lower surface of the inner ring plate 3. Such a layout ensures that the sensor can fully cover the possible peeling defect area, and improves the comprehensiveness and accuracy of the detection. The control component is the center of the system, and its data end is electrically connected to the data end of the external voltage sensor assembly 4 and the embedded voltage sensor assembly 5. This connection mode enables the control component to receive data from the sensor in real time and analyze and process it. In this way, the system can quickly respond and accurately identify the peeling defects of the internal interface of the steel tube concrete component.

[0015] Simply put, the steel tube concrete interface peeling defect positioning detection system installs piezoelectric ceramic sensors or embedded voltage sensors on the outer wall of the steel component module 1 and the lower surface of the inner ring plate 3 of the steel and concrete composite component 10. At least 3 groups of sensors should be arranged under the same boundary conditions. After the piezoelectric ceramic sensor is installed, a multimeter is used to detect its insulation performance. Next, an impedance analyzer of the control component is used to apply alternating voltage excitation to the piezoelectric ceramic sensor. Using the forward and reverse piezoelectric effects of the piezoelectric ceramic material, the piezoelectric ceramic sensor outputs an electromechanically coupled admittance frequency curve. Finally, the admittance data is transmitted to the computer for data processing and converted into image data; a convolutional neural network model is established to locate the defects.

[0016] refer to Figure 4 Preferably, the embedded voltage sensor component 5 includes a first piezoelectric material module, a first shielded wire, a first insulating coating, a first waterproof coating, and an encapsulated concrete block 11, wherein the positive and negative electrodes of the first piezoelectric material module are respectively welded to the welding ends of the first shielded wire, the welding ends are coated with a first insulating coating and a first waterproof coating, the first piezoelectric material module and the welding ends of the first shielded wire are encapsulated in the encapsulated concrete block, and the second end of the first shielded wire is electrically connected to the data end of the control component.

[0017] In this embodiment, the embedded voltage sensor assembly 5 includes a piezoelectric material module 16 positive and negative electrodes and a shielded wire 6 welded thereto respectively. After the positive and negative electrodes of the shielded wire 6 are welded to the positive and negative electrodes of the piezoelectric material module 16 respectively, a waterproof layer 13 and an insulating layer 12 need to be coated respectively to ensure the insulation performance of the embedded voltage sensor assembly 5. A multimeter is required to detect its insulation performance. Finally, the piezoelectric material module 16 is encapsulated into the encapsulated concrete block 11 to form an embedded voltage sensor assembly 5. The piezoelectric material module 16 includes materials having a piezoelectric effect, such as piezoelectric ceramic sheets and piezoelectric films. It should be noted that the piezoelectric material module 16, shielded wire 6, waterproof layer 13 and insulating layer 12 included in the external voltage sensor assembly 4 are consistent with the embedded voltage sensor assembly 5.

[0018] refer to Figure 5 Preferably, the external voltage sensor assembly 5 includes a healthy position voltage ceramic sensor module 14 and a damaged position voltage ceramic sensor module 15, wherein the damaged position voltage ceramic sensor module 15 is configured in an interface debonding damage area preset on the outer wall surface of the steel component module 1, and the healthy position voltage ceramic sensor module 14 is configured in a healthy area preset on the outer wall surface of the steel component module 1.

[0019] Preferably, the healthy position voltage ceramic sensor module 14 includes a second piezoelectric material module, a second shielded wire, a second insulating coating, and a second waterproof coating. The positive and negative electrodes of the second piezoelectric material module are respectively welded to the welding ends of the second shielded wire, and the welding ends are coated with a second insulating coating and a second waterproof coating. The second end of the second shielded wire is electrically connected to the data end of the control component.

[0020] Preferably, the damage location voltage ceramic sensor module 15 includes a third piezoelectric material module, a third shielded wire, a third insulating coating, and a third waterproof coating. The positive and negative electrodes of the third piezoelectric material module are respectively welded to the welding ends of the third shielded wire, and the welding ends are coated with a third insulating coating and a third waterproof coating. The second end of the third shielded wire is electrically connected to the data end of the control component.

[0021] In this embodiment, after the piezoelectric ceramic sheet and the positive and negative electrodes of the shielded wire 6 are welded together, the external piezoelectric sensor needs to be coated with a waterproof layer 13 and an insulating layer 12, and then a multimeter is used to detect its insulation performance; the surface of the steel component at the test position of the steel tube concrete component to be tested is polished and smoothed with a grinder or sandpaper; a cotton ball is dipped in an appropriate amount of anhydrous ethanol, and the polished position is wiped and cleaned, and the place is kept clean; and a layer of insulating layer is evenly applied to the part, and after the insulating layer solidifies, a multimeter is used to detect its insulation performance; after confirming the insulation performance of the insulating layer 12, the external piezoelectric sensor is attached to the outer wall of the steel component module 1 or the lower surface of the inner ring plate 3 of the steel tube at the intended test position with insulating glue. In actual use, the user can number the piezoelectric sensors installed on the outer surface of the steel and concrete composite component 10, such as Figure 6 As shown, this number will correspond to the number of the damage indicator of each piezoelectric sensor in the subsequent steps.

[0022] It should be noted that the second piezoelectric material module and the third piezoelectric material module include materials having piezoelectric effect such as piezoelectric ceramic sheets and piezoelectric films, which are similar to the second piezoelectric material module and will not be described in detail here.

[0023] Preferably, the first insulating coating, the second insulating coating, the third insulating coating, the first waterproof coating, the second waterproof coating, and the third waterproof coating are all epoxy resin adhesive layers or insulating adhesive layers, and the adhesive layer thickness is 0.05 mm-0.15 mm.

[0024] In this embodiment, the thickness of the adhesive layer is preferably 0.1 mm. The selection of this thickness not only ensures good bonding between the sensor and the steel tube concrete component, but also reduces the impact of the adhesive layer on the detection results, thereby improving the stability and reliability of the system. This precise coating thickness control and strict installation process not only improve the installation quality of the sensor, but also enhances the overall performance of the system. In this way, the sensor can more accurately capture subtle changes in the internal interface of the steel tube concrete component, thereby achieving early warning and precise positioning of peeling defects. In addition, this design also helps to reduce the impact of external environmental factors on sensor performance, ensuring the accuracy and consistency of detection data.

[0025] Preferably, the control component includes a data transmission line 7, an impedance analyzer 8, and a controller 9. The input end of the impedance analyzer 8 is electrically connected to the shielded wire 6 of the external voltage sensor component 4 and the shielded wire 6 of the embedded voltage sensor component 5. The impedance analyzer 8 is electrically connected to the controller 9 through the data transmission line 7.

[0026] In this embodiment, the impedance analyzer 8 is used to select multiple test frequency bands and call output parameters for external piezoelectric sensors and embedded voltage sensors. Specifically, the impedance analyzer 8 is used to apply a constant amplitude AC voltage to the piezoelectric sensor, and at the same time, the impedance analyzer is used to collect the admittance modulus value of the electromechanical coupling system composed of the external piezoelectric sensor, steel pipe, bonding interface and concrete. The frequency band selected by the impedance analyzer 8 is 1 kHz-1000kHz, and the number of sampling points is 800.

[0027] See also Figure 2 The following is an example of attaching an external piezoelectric sensor to the outer wall of a steel component module 1 of a 580 mm×580 mm×1000 mm square steel tube concrete component 10, attaching an embedded sensor to the lower surface of its inner ring plate 3, and detecting the debonding defect at the inner wall of the steel tube. The arrangement of the external piezoelectric sensor on the outer wall of the steel component module 1 is shown in FIG. Figure 6 To test the effectiveness of the proposed method, before pouring concrete in the steel-concrete composite member 10, Figure 6 A thin wooden board is pasted on the inner wall of the steel component module 1 corresponding to the dotted area shown to simulate an artificial interface peeling defect. Among them, the external piezoelectric sensors 18, 19, 20, 21 and 22 are in the interface peeling damage area, while the external piezoelectric sensors 14, 15, 16 and 17 are in the healthy area. The admittance modulus values ​​of the external piezoelectric sensors 14 to 17 at the healthy position are selected as references, and the admittance signal of the external piezoelectric sensor 12 is selected as the signal at the healthy position.

[0028] The control component is configured to implement the following steps by executing a computer program stored therein: S1, acquiring detection data collected by the external voltage sensor component 4 and the embedded voltage sensor component 5, and performing Euclidean calculation preprocessing on the detection data; Specifically, step S1 includes: acquiring detection data collected by the external voltage sensor component 4 and the embedded voltage sensor component 5; Performing Euclidean calculation on the detection data to generate an ED matrix; The ED matrix is ​​normalized, and the normalized ED matrix is ​​converted into an RGB format to obtain an input feature map.

[0029] In this embodiment, the output signal of the piezoelectric ceramic sensor at the defective position and the output signal of the piezoelectric ceramic sensor at the healthy position are Euclideanly calculated to form an ED matrix, and the raw data is converted into RGB format after normalization and used as the input layer of the convolutional neural network.

[0030] S2, input the preprocessed detection data into the preset convolutional neural network model to obtain the defect location result.

[0031] Specifically, step S2 includes: using the input feature map as an input of a convolutional neural network model, and the convolution layer of the model performs convolution operation processing on the input feature map to extract two-dimensional features; The pooling layer of the convolutional neural network model downsamples the input feature map after the convolution operation to reduce the dimension of the input feature map and extract secondary two-dimensional features; Repeat the above steps until the end of the convolutional neural network model is connected to a fully connected layer to map the extracted two-dimensional features into a one-dimensional vector; The one-dimensional vector is classified by a Softmax classifier, the damage state is output, and the location of the defect is determined. The calculation formula is: , , where l is the number of layers, w is the convolution kernel, is the mth neuron in the lth layer, is the mth bias of the lth layer, is the mth output of the lth layer, f(x) is the pooling operation, and k is the total number of layers of the convolutional neural network model.

[0032] In this embodiment, the convolution kernel is used to perform convolution operation on the upper feature map to extract features. The pooling layer downsamples the feature map after the convolution layer to achieve the effect of reducing the dimension and extracting secondary features. After passing through multiple convolutional layers and pooling layers, the network end is connected to a fully connected layer to map the two-dimensional features into a one-dimensional vector, which is passed to the Softmax classifier to classify the extracted features. The output layer is the damage state, thereby determining the location of the defect.

[0033] It should be noted that before using the convolutional neural network model, the test data is input into the network model for training to obtain the interface bonding state classification of each piezoelectric ceramic sensor location, and then the interface bonding state between the inner wall of the steel pipe and the core concrete, and the inner ring plate and the concrete at the location of each piezoelectric ceramic sensor is obtained to determine whether there is an interface peeling defect. When the generated result reaches the preset value, the trained convolutional neural network model is obtained.

[0034] In actual operation, external and embedded voltage sensors generate electromechanical coupled admittance frequency curves by applying alternating voltage excitation and utilizing the forward and reverse piezoelectric effects of piezoelectric ceramic materials. These curves are converted into image data and input into the convolutional neural network model. The model extracts and classifies features through convolutional layers, pooling layers, and fully connected layers, and finally determines the location of the defect through the Softmax classifier. This process not only improves the efficiency of detection, but also improves the accuracy of detection through deep learning algorithms.

[0035] Specifically, taking the above-mentioned 580mm×580mm×1000mm square steel tube concrete component as an example, the admittance frequency data (i.e. Figure 7 data) into RGB image data, such as Figure 8 As shown in the figure, a convolutional neural network is established, and the training set data is iteratively learned. The test set is used to verify the recognition effect of the model, effectively identifying the debonding defect at the steel tube concrete interface and locating the location where the defect exists. The real damage label (zheng) is completely consistent with the test result (yuce), and the accuracy of the test result is 100%. Fig. 9 The results of five tests are shown in Figure 1. The debonding defect at the steel tube concrete interface is effectively identified and the damage position is located, which is consistent with the actual situation.

[0036] In summary, in this embodiment, the steel tube concrete interface peeling defect positioning detection system is respectively installed with an external piezoelectric sensor assembly 4 and an embedded piezoelectric sensor assembly 5 on the outer surface of the steel component module 1 and the lower surface of the inner ring plate 3 of the steel and concrete composite component 10. The impedance analyzer 8 is used to apply alternating voltage excitation to the piezoelectric sensor. First, at least two frequency bands are selected for preliminary testing, and then a suitable frequency band is determined as the detection frequency band, and the number of frequency sampling points on each frequency band is not less than 400. The MATLAB software is used to pre-process the output admittance signal of the piezoelectric ceramic sensor, the admittance-frequency data obtained by the detection is Euclidean calculated, and the normalized matrix is ​​converted into RGB format data; data enhancement is used to expand the RGB format data, increase samples, and reduce the risk of overfitting in the data training process. 10% of the processed RGB data is used as a test set, and the rest is used as a training set. Finally, the python torch is used to establish a convolutional neural network model to judge the interface peeling defects of each measuring point and realize the peeling defect positioning; among them, neural networks have gradually become a practical and accurate method for analyzing and collecting data using data-driven methods. CNN has the ability to accurately locate defects by learning complex data patterns. A convolutional neural network is established to iteratively learn the data of the training set, and the recognition effect of the model is tested using the test set. It can effectively identify the debonding defects at the steel tube concrete interface and accurately locate the defect location.

[0037] Compared with the existing solutions, the steel tube concrete interface peeling defect positioning detection system installs an embedded piezoelectric sensor under the inner ring plate for piezoelectric admittance detection, and uses the convolutional network algorithm to detect and locate the interface defects between the lower surface of the inner ring plate and the concrete. At the same time, the use of convolutional neural networks can achieve efficient processing and real-time analysis of a large amount of collected data, thereby quickly locating internal defects. The steel tube concrete interface peeling defect positioning detection system provides a fast and accurate method for real-time detection of interface peeling of steel tube concrete components, especially the piezoelectric admittance in the high-frequency range has a high sensitivity to local initial damage. The system can quickly detect and accurately determine the defect location, and is easy to operate, fast response, and high visibility.

[0038] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A steel tube concrete interface peeling defect positioning and detection system, characterized in that: include: A control component, a steel and concrete composite component, an inner ring plate, an external voltage sensor component, and an embedded voltage sensor component, wherein the inner ring plate is arranged inside the steel and concrete composite component, the external voltage sensor component is arranged on the outer wall surface of the steel and concrete composite component, the embedded voltage sensor component is arranged on the lower surface of the inner ring plate, and the data end of the control component is electrically connected to the data end of the external voltage sensor component and the data end of the embedded voltage sensor component; The control component is configured to implement the following steps by executing a computer program stored therein: Acquire detection data collected by the external voltage sensor component and the embedded voltage sensor component, and perform Euclidean calculation preprocessing on the detection data; The preprocessed detection data is input into the preset convolutional neural network model to obtain the defect location results.

2. The steel tube concrete interface debonding defect positioning and detection system according to claim 1 is characterized in that: The steel and concrete composite component includes a steel component module and concrete, and the concrete is poured inside the steel component module.

3. The steel tube concrete interface debonding defect positioning and detection system according to claim 2 is characterized in that: The embedded voltage sensor component includes a first piezoelectric material module, a first shielded wire, a first insulating coating, a first waterproof coating, and an encapsulated concrete block, wherein the positive and negative electrodes of the first piezoelectric material module are respectively welded to the welding ends of the first shielded wire, the welding ends are coated with a first insulating coating and a first waterproof coating, the first piezoelectric material module and the welding ends of the first shielded wire are encapsulated in the encapsulated concrete block, and the second end of the first shielded wire is electrically connected to the data end of the control component.

4. The steel tube concrete interface debonding defect positioning and detection system according to claim 3 is characterized in that: The external voltage sensor assembly includes a healthy position voltage ceramic sensor module and a damaged position voltage ceramic sensor module, wherein the damaged position voltage ceramic sensor module is configured at a position on the outer wall of the steel pipe corresponding to an interface peeling damage area preset on the inner wall of the steel component module, and the healthy position voltage ceramic sensor module is configured in a healthy area preset on the outer wall surface of the steel component module.

5. The steel tube concrete interface debonding defect positioning and detection system according to claim 4 is characterized in that: The healthy position voltage ceramic sensor module includes a second piezoelectric material module, a second shielded wire, a second insulating coating, and a second waterproof coating. The positive and negative electrodes of the second piezoelectric material module are respectively welded to the welding ends of the second shielded wire, and the welding ends are coated with a second insulating coating and a second waterproof coating. The second end of the second shielded wire is electrically connected to the data end of the control component.

6. The steel tube concrete interface debonding defect positioning and detection system according to claim 5 is characterized in that: The damage location voltage ceramic sensor module includes a third piezoelectric material module, a third shielded wire, a third insulating coating, and a third waterproof coating. The positive and negative electrodes of the third piezoelectric material module are respectively welded to the welding ends of the third shielded wire, and the welding ends are coated with a third insulating coating and a third waterproof coating. The second end of the third shielded wire is electrically connected to the data end of the control component.

7. The steel tube concrete interface debonding defect positioning and detection system according to claim 6 is characterized in that: The first insulating coating, the second insulating coating, the third insulating coating, the first waterproof coating, the second waterproof coating, and the third waterproof coating are all epoxy resin adhesive layers or insulating adhesive layers, and the thickness of the adhesive layer is 0.05 mm-0.15 mm.

8. The steel tube concrete interface debonding defect positioning and detection system according to claim 1 is characterized in that: The control component includes a data transmission line, an impedance analyzer, and a controller. The input end of the impedance analyzer is electrically connected to the shielded wire of the external voltage sensor component and the shielded wire of the embedded voltage sensor component. The impedance analyzer is electrically connected to the controller through the data transmission line.

9. The steel tube concrete interface debonding defect positioning and detection system according to claim 1, characterized in that: Acquire the detection data collected by the external voltage sensor component and the embedded voltage sensor component, and perform Euclidean calculation preprocessing on the detection data, specifically: Acquiring detection data collected by the external voltage sensor component and the embedded voltage sensor component; Performing Euclidean calculation on the detection data to generate an ED matrix; The ED matrix is ​​normalized, and the normalized ED matrix is ​​converted into an RGB format to obtain an input feature map.

10. The steel tube concrete interface debonding defect positioning and detection system according to claim 9, characterized in that: The preprocessed detection data is input into the preset convolutional neural network model to obtain the defect location results, specifically: The input feature map is used as an input of a convolutional neural network model, and the convolution layer of the model performs convolution operation processing on the input feature map to extract two-dimensional features; The pooling layer of the convolutional neural network model downsamples the input feature map after the convolution operation to reduce the dimension of the input feature map and extract secondary two-dimensional features; Repeat the above steps until the end of the convolutional neural network model is connected to a fully connected layer to map the extracted two-dimensional features into a one-dimensional vector; The one-dimensional vector is classified by a Softmax classifier, the damage state is output, and the location of the defect is determined. The calculation formula is: , , where l is the number of layers, w is the convolution kernel, is the mth neuron in the lth layer, is the mth bias of the lth layer, is the mth output of the lth layer, f(x) is the pooling operation, and k is the total number of layers of the convolutional neural network model.

Images