Steel profile nondestructive testing device and testing method thereof

Abstract

The invention discloses a nondestructive testing device for a steel profile and a testing method of the nondestructive testing device. According to the method disclosed by the invention, the ultrasonic technology is combined with optical imaging, so that synchronous detection of surface and internal defects can be completed at one time. Ultrasonic waves penetrate through the interior of the steel profile, and hidden defects such as cracks and air holes are accurately recognized; the optical camera module captures visible problems such as surface scratches and corrosion in real time. The integrated detection mode avoids the tedious process of step-by-step operation required by a traditional method, for example, magnetic powder detection of surface defects and ultrasonic scanning of an internal structure are not needed, the detection period is greatly shortened, and the method is especially suitable for steel structure buildings, bridge engineering and other scenes with high requirements for detection efficiency and has good application prospects. The full-process automatic design effectively reduces the influence of human intervention, and ensures the scanning continuity of the detection area; the intelligent algorithm automatically analyzes the ultrasonic echo signal and the optical image data, and judges the defect level according to the preset standard, thereby avoiding subjective misjudgment of manual visual inspection.

Description

Technical Field

[0001] This invention belongs to the field of steel profile testing technology, specifically a non-destructive testing device and testing method for steel profiles. Background Technology

[0002] Steel profiles, as a fundamental material in modern industry, are widely used in construction, transportation, energy, and other fields. With increasing demands on material performance in engineering structures, the quality inspection of steel profiles has become a crucial link in ensuring engineering safety. Currently, non-destructive testing (NDT) technologies mainly include ultrasonic testing, magnetic particle testing, penetrant testing, eddy current testing, and radiographic testing. These technologies each have their own focus in industrial applications, but they generally suffer from low testing efficiency, complex equipment, and stringent environmental requirements. In recent years, with the development of intelligent manufacturing, automated NDT equipment has gradually become a research hotspot, but many technical bottlenecks still exist in the online inspection of steel profiles.

[0003] In existing technologies, the following methods are mainly used for non-destructive testing of steel profiles: The main body of the detection device (1) Ultrasonic detection: internal defects are detected by emitting ultrasonic waves through the probe and receiving the reflected signals. The advantage is that the detection depth is large, but it is sensitive to surface roughness and requires a coupling agent. Conveying mechanism (2) Magnetic particle inspection: Utilizes magnetic field and magnetic powder to display surface and near-surface defects. It is simple to operate but only applicable to ferromagnetic materials; Detection unit (3) Eddy current detection: Based on the principle of electromagnetic induction, it detects surface defects. It is fast but difficult to detect deep defects. Control module (4) X-ray detection: It detects internal defects through X-ray or γ-ray imaging. The results are intuitive but there are radiation safety hazards.

[0004] However, the main drawbacks of existing technologies include: ultrasonic testing requires manual coupling and is inefficient; magnetic particle testing is only applicable to surface defects; eddy current testing requires material conductivity and cannot detect internal defects; and X-ray inspection equipment is expensive and poses radiation risks. Furthermore, existing technologies struggle to achieve automated online inspection of steel profiles, failing to meet the high-efficiency demands of modern industrial production. Summary of the Invention

[0005] The purpose of this invention is to provide a non-destructive testing device and method for steel profiles in order to solve the problems mentioned above.

[0006] The technical solution adopted in this invention is as follows: a non-destructive testing device for steel profiles, the device comprising: a testing device body (1), a conveying mechanism (2), a testing unit (3), and a control module (4); The main body of the detection device (1) is supported by a frame structure, which includes a base, columns and beams; The conveying mechanism (2) includes a conveying roller and a drive motor. The conveying roller is mounted on the base, and the drive motor is connected to the conveying roller via a chain. The detection unit (3) includes an ultrasonic probe array and an optical camera module. The ultrasonic probe array is installed below the crossbeam via an adjustable bracket, and the optical camera module is fixed in the middle of the crossbeam. The control module (4) includes a data processing unit and a human-machine interface, which are installed in the control box on the side of the device.

[0007] In a preferred embodiment, the ultrasonic probe array uses 8 wideband probes ranging from 0.5MHz to 10MHz.

[0008] In a preferred embodiment, the optical camera module employs a 5-megapixel industrial camera.

[0009] In a preferred embodiment, the conveyor rollers are made of 304 stainless steel and covered with a polyurethane anti-slip layer; the drive motor is a 1.5kW servo motor.

[0010] In a preferred embodiment, a non-destructive testing method for steel profiles includes the following steps: S1: Placement and positioning of steel profiles: Place the steel profile to be inspected at the beginning of the conveyor roller, ensuring that it is aligned with the roller, and fix its position with the anti-slip layer to prevent it from shifting during the inspection process; S2: Detection parameter settings: Input detection parameters through the human-computer interaction interface, including material type, detection speed, ultrasonic frequency range and imaging resolution of the optical camera module, and simultaneously select whether to enable the laser ranging module to monitor thickness changes; S3: Automated Synchronous Detection After the detection program is started, the conveyor rollers drive the steel profile into the detection area at a constant speed; the ultrasonic probe array scans internal defects through the principle of sound wave reflection, the optical camera module captures surface defects based on machine vision, and if the optimization scheme is enabled, the laser ranging module records the profile thickness data in real time. S4: Data Fusion and Defect Identification The control module receives ultrasonic signals and optical image data, and through algorithm fusion analysis, automatically identifies the defect type, location and size, determines whether it is qualified based on preset standards, and generates a defect map containing three-dimensional positioning information. S5: Report Output and Workpiece Sorting After the inspection is completed, the system automatically generates a digital inspection report and sends it to the local terminal or cloud platform through the control module; the inspected workpieces are output to the designated area by the conveyor rollers, and the unqualified products are automatically marked or sorted.

[0011] In a preferred embodiment, in step S1, the steel profile needs to be placed horizontally at the starting end of the conveyor roller. The roller surface is made of 304 stainless steel and covered with a polyurethane anti-slip layer, with a friction coefficient of not less than 0.8, to ensure that there is no slippage or offset during workpiece transmission. The positioning accuracy requires that the longitudinal alignment error is less than ±1.5mm and the lateral offset does not exceed 5% of the roller width. The drive motor is a 1.5kW servo motor, and the speed fluctuation rate is less than 0.5% through a closed-loop control system to ensure the stability of feeding in the subsequent detection area. The workpiece length is suitable for a range of 2m to 12m, and the maximum load capacity is 3 tons.

[0012] In a preferred embodiment, in step S2, the material type needs to be selected from a preset database through a human-machine interface, covering 20 common steel grades such as Q235 and Q345. The system automatically loads the corresponding sound velocity and density physical parameters. The detection speed is adjustable from 0.5 to 2 m / min, with a recommended standard speed of 1 m / min to balance efficiency and accuracy. The ultrasonic probe array is configured with 8 independent channels, covering a frequency range of 0.5 MHz to 10 MHz, where the low-frequency band is used to detect deep defects, and the high-frequency band captures near-surface microcracks. The optical camera module uses a 5-megapixel global shutter industrial camera, with the frame rate matching the transmission speed to ensure that at least 10 effective pixels are collected per millimeter of workpiece surface. If the thickness monitoring function is enabled, the laser ranging module needs to be preheated and calibrated before the workpiece enters the detection area, and the measurement accuracy needs to reach ±0.05 mm.

[0013] In a preferred embodiment, in step S3, the conveyor rollers drive the workpiece into the detection area at a constant linear speed, with the speed deviation controlled within ±0.1m / min; the ultrasonic probe array adopts a matrix layout, with an adjacent probe spacing of 50mm, and the sound beam incident angle is automatically adjusted according to the profile cross-sectional shape to ensure 100% sound wave coverage; the pulse repetition frequency of each probe is set to 5kHz, and the sampling depth is matched to more than twice the maximum thickness of the profile; the optical detection system uses coaxial light source illumination, with the light intensity set to be adjustable from 2000 to 5000 lux, and the exposure time is dynamically optimized according to the surface reflectivity to ensure that the image grayscale value distribution is within the effective range of 80-220; when thickness monitoring is enabled, the laser line scanning frequency must be synchronized with the conveying speed, and no less than 3 thickness data points are collected per millimeter of length.

[0014] In a preferred embodiment, in step S4, the acoustic wave reflection signal acquired by the ultrasonic probe array is first analyzed in the time-frequency domain, and the defect echo features are extracted using an improved wavelet packet decomposition algorithm; a multi-modal feature joint discrimination model is constructed; the ultrasonic feature vector and the image feature vector are input into a dual-channel neural network, and feature interaction is achieved through a cross-attention mechanism; a gated fusion unit is used to dynamically adjust the contribution weights of each mode; for the generation of three-dimensional positioning information, a spatial positioning algorithm based on the acoustic wave propagation time difference is proposed; through the collaborative measurement of multiple probe arrays, a set of geometric constraint equations for the acoustic wave propagation path is established, and an improved particle swarm optimization algorithm is used to quickly solve for the spatial coordinates of the defect; finally, through the defect size quantization module, combined with the acoustic wave attenuation coefficient and the image pixel resolution, a composite calculation model of the defect volume is established to achieve sub-millimeter level precision defect quantification assessment; The formula for the multimodal feature fusion weight function is as follows: Where Fu and Fv represent ultrasonic and visual feature vectors, respectively, Wu and Wv are learnable projection matrices, and α and β are modal confidence coefficients; this function effectively solves the modal interference problem of the traditional weighted average method by dynamically evaluating the information entropy density of each modal feature and adaptively adjusting the fusion weights. The formula for the sound wave propagation path optimization equation is: In the formula: di represents the Euclidean distance from the i-th probe to the defect point P, v is the sound velocity of the material, ti is the measured propagation time, and λ is the spatial continuity constraint coefficient. This equation introduces a spatial gradient regularization term on the basis of the traditional least squares method, which effectively suppresses the positioning jump caused by noise and realizes stable three-dimensional reconstruction of internal defects of complex structural components.

[0015] In a preferred embodiment, in step S5, the inspection report is automatically generated in PDF format, including defect type code, three-dimensional coordinate positioning data and equivalent defect size calculation value, with coordinate positioning accuracy reaching ±0.3mm.

[0016] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. In this invention, by combining ultrasonic technology with optical imaging, the method can simultaneously detect surface and internal defects in one operation. Ultrasonic waves penetrate the interior of the steel profile, accurately identifying hidden defects such as cracks and pores; the optical camera module captures visible problems such as surface scratches and corrosion in real time. This integrated detection mode avoids the cumbersome step-by-step process required by traditional methods. For example, it eliminates the need to first perform magnetic particle testing of surface defects and then use ultrasonic scanning of the internal structure, significantly shortening the detection cycle. It is particularly suitable for scenarios with high detection efficiency requirements, such as steel structure buildings and bridge engineering.

[0017] 2. In this invention, the fully automated design effectively reduces the impact of human intervention. The conveyor rollers are precisely controlled by servo motors to move the steel profiles at a uniform speed, with a speed deviation of less than ±0.1 m / min, ensuring the continuity of scanning in the detection area. Intelligent algorithms automatically analyze ultrasonic echo signals and optical image data, determining the defect level according to preset standards, avoiding subjective misjudgments from manual visual inspection. For example, for the identification of micro-cracks, the system automatically classifies them using feature thresholds, significantly improving the consistency of detection results when different operators use the same equipment.

[0018] 3. In this invention, the modular design allows the equipment to flexibly adapt to steel profiles of different specifications. The spacing of the probe array can be adjusted manually or electrically via a mechanical structure, covering the testing needs from narrow angle steel to wide H-beams; the anti-slip layer and drive parameters of the conveyor rollers can be dynamically optimized according to the weight and length of the workpiece. For example, when testing H-beams, only the probe layout and conveying speed need to be adjusted to complete the adaptation, without replacing core components, significantly reducing equipment modification costs and time.

[0019] 4. In this invention, the non-contact testing technology avoids direct contact with the steel profile surface throughout the entire process. The optical camera module uses long-distance, high-resolution imaging, eliminating the need for developing agents or coupling agents; the ultrasonic probe achieves non-contact scanning through air coupling or electromagnetic induction, completely solving the problems of magnetic suspension residue in traditional magnetic particle testing or chemical reagent contamination in penetrant testing. This method is particularly suitable for precision profiles with polished or coated surfaces, ensuring that the testing process does not cause secondary damage to the workpiece.

[0020] 5. In this invention, the integrated data processing system automatically converts inspection results into standardized reports, including the three-dimensional coordinates, equivalent dimensions, and type classification of defect locations, and supports local storage and cloud synchronization. Managers can quickly retrieve historical data through keyword searches, such as filtering records by time, material batch, or defect type, simplifying the quality traceability process. Furthermore, the cloud backup function facilitates cross-departmental collaborative analysis; for example, engineering teams can directly access inspection data to assess structural safety without relying on paper documents or manual summarization, significantly improving management efficiency.

[0021] 6. In this invention, an ultrasonic probe array combined with a wavelet packet decomposition algorithm accurately captures the acoustic reflection characteristics of defects such as internal cracks and pores. The optical camera module uses a lightweight convolutional neural network to analyze image features such as surface scratches and corrosion in real time, simultaneously detecting internal and external defects, avoiding the efficiency loss caused by step-by-step operations in traditional methods. A multimodal feature fusion algorithm is adopted to dynamically adjust the weights of ultrasonic and optical data, and a cross-attention mechanism is used to eliminate the risk of misjudgment from a single detection method. For example, the correlation between surface scratches and internal cracks is automatically analyzed through a dual-channel neural network, improving the accuracy of defect classification. Adjustable probe spacing and dynamic acoustic beam incident angle design, combined with adaptive adjustment of conveyor roller parameters, enable rapid switching detection from I-beams to H-beams.

[0022] 7. In this invention, the system automatically matches the probe layout and scanning path for different cross-sectional profiles, eliminating the need for manual recalibration. Non-contact optical imaging combined with air-coupled ultrasonic technology avoids the chemical contamination or ultrasonic coupling agent residue problems associated with traditional magnetic particle testing, making it particularly suitable for non-destructive testing of surface-coated or precision-machined profiles. Based on an improved particle swarm optimization algorithm, the three-dimensional defect localization technology generates a digital report with coordinates, with a defect location error of less than 0.3 mm. The cloud synchronization function supports real-time access across multiple terminals, simplifying the quality traceability process; for example, it allows for quick location of the inspection records for a specific batch of profiles using timestamps. Attached Figure Description

[0023] Figure 1 This is an overall system block diagram of the present invention; Figure 2 This is a flowchart of the detection method in this invention.

[0024] The diagram shows: 1 - main body of the detection device, 2 - conveying mechanism, 3 - detection unit, and 4 - control module. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0026] Reference Figure 1-2 A non-destructive testing device for steel profiles, comprising: a main body of the testing device 1, a conveying mechanism 2, a testing unit 3, and a control module 4; The main body of the detection device 1 is supported by a frame structure, which includes a base, columns and beams; The conveying mechanism 2 includes a conveyor roller and a drive motor. The conveyor roller is mounted on the base, and the drive motor is connected to the conveyor roller via a chain. The detection unit 3 includes an ultrasonic probe array and an optical camera module. The ultrasonic probe array is mounted below the crossbeam via an adjustable bracket, and the optical camera module is fixed in the middle of the crossbeam. The control module 4 includes a data processing unit and a human-machine interface, and is installed in the control box on the side of the device.

[0027] The ultrasonic probe array uses 8 wideband probes ranging from 0.5MHz to 10MHz.

[0028] The optical camera module uses a 5-megapixel industrial camera.

[0029] The conveyor rollers are made of 304 stainless steel and covered with a polyurethane anti-slip layer; the drive motor is a 1.5kW servo motor.

[0030] A non-destructive testing method for steel profiles, comprising the following steps: S1: Placement and positioning of steel profiles: Place the steel profile to be tested at the beginning of the conveyor roller, ensuring it is aligned with the roller, and fix its position with the anti-slip layer to prevent it from shifting during the testing process.

[0031] S2: Detection parameter settings: Input detection parameters through the human-computer interaction interface, including material type (such as Q235, Q345, etc.), detection speed (0.5-2m / min), ultrasonic frequency range (0.5MHz-10MHz), and imaging resolution of the optical camera module (5 million pixels). Simultaneously select whether to enable the laser ranging module (optimized scheme) to monitor thickness changes.

[0032] S3: Automated Synchronous Detection After the detection program is started, the conveyor rollers drive the steel profile into the detection area at a constant speed; the ultrasonic probe array scans internal defects (such as cracks and pores) through the principle of sound wave reflection, and the optical camera module captures surface defects (such as scratches and rust) based on machine vision. If the optimization scheme is enabled, the laser ranging module records the profile thickness data in real time.

[0033] S4: Data Fusion and Defect Identification The control module receives ultrasonic signals and optical image data, and through algorithmic fusion analysis (such as signal threshold judgment and image feature extraction), it automatically identifies the defect type, location and size, determines whether it is qualified based on preset standards, and generates a defect map containing three-dimensional positioning information.

[0034] S5: Report Output and Workpiece Sorting After the inspection is completed, the system automatically generates a digital inspection report (including defect distribution map, inspection parameters and judgment results) and sends it to the local terminal or cloud platform (optimized solution) through the control module; the inspected workpieces are output to the designated area by the conveyor roller, and the unqualified products are automatically marked or sorted.

[0035] In step S1, the steel profile must be placed horizontally at the starting end of the conveyor rollers. The roller surface is made of 304 stainless steel and covered with a polyurethane anti-slip layer, with a friction coefficient of not less than 0.8, ensuring no slippage or offset during workpiece transport. Positioning accuracy requires a longitudinal alignment error of less than ±1.5mm and a lateral offset not exceeding 5% of the roller width. A 1.5kW servo motor is used as the drive motor, and a closed-loop control system ensures a speed fluctuation rate of less than 0.5%, guaranteeing the stability of feeding in the subsequent inspection area. The workpiece length range is 2m to 12m, and the maximum load capacity is 3 tons, meeting the specifications of common H-beams, I-beams, and other profiles.

[0036] In step S2, the material type needs to be selected from a pre-set database through the human-machine interface, covering 20 common steel grades such as Q235 and Q345. The system automatically loads the corresponding physical parameters such as sound velocity and density. The detection speed is adjustable from 0.5 to 2 m / min, with a recommended standard speed of 1 m / min to balance efficiency and accuracy. The ultrasonic probe array is configured with 8 independent channels, covering a frequency range of 0.5 MHz to 10 MHz. The low-frequency band (0.5-2 MHz) is used to detect deep defects, while the high-frequency band (5-10 MHz) captures near-surface microcracks. The optical camera module uses a 5-megapixel global shutter industrial camera, with a frame rate matching the transmission speed to ensure that at least 10 effective pixels are collected per millimeter of workpiece surface. If the thickness monitoring function is enabled, the laser ranging module needs to be preheated and calibrated before the workpiece enters the detection area, and the measurement accuracy needs to reach ±0.05 mm.

[0037] In step S3, the conveyor rollers drive the workpiece into the inspection area at a constant linear speed, with the speed deviation controlled within ±0.1 m / min. The ultrasonic probe array adopts a matrix layout, with an adjacent probe spacing of 50 mm. The sound beam incident angle is automatically adjusted according to the profile cross-sectional shape to ensure 100% sound wave coverage. The pulse repetition frequency of each probe is set to 5 kHz, and the sampling depth is matched to more than twice the maximum thickness of the profile. The optical inspection system uses coaxial light source illumination, with the light intensity set to be adjustable from 2000 to 5000 lux. The exposure time is dynamically optimized according to the surface reflectivity to ensure that the image grayscale value distribution is within the effective range of 80-220. When thickness monitoring is enabled, the laser line scanning frequency must be synchronized with the conveyor speed, and no less than 3 thickness data points must be collected per millimeter of length.

[0038] In step S4, the acoustic wave reflection signal acquired by the ultrasonic probe array is first analyzed in the time-frequency domain, and the defect echo features are extracted using an improved wavelet packet decomposition algorithm. A multi-modal feature joint discrimination model is constructed. The ultrasonic feature vector and the image feature vector are input into a dual-channel neural network, and feature interaction is achieved through a cross-attention mechanism. A gated fusion unit is used to dynamically adjust the contribution weights of each mode. For the generation of three-dimensional positioning information, a spatial positioning algorithm based on the acoustic wave propagation time difference is proposed. Through the collaborative measurement of multiple probe arrays, a set of geometric constraint equations for the acoustic wave propagation path is established, and an improved particle swarm optimization algorithm is used to quickly solve for the spatial coordinates of the defect. Finally, through the defect size quantization module, combined with the acoustic wave attenuation coefficient and image pixel resolution, a composite calculation model of the defect volume is established to achieve sub-millimeter-level precision defect quantification assessment. The formula for the multimodal feature fusion weight function is as follows: Where Fu and Fv represent the ultrasonic and visual feature vectors, respectively, Wu and Wv are learnable projection matrices, and α and β are modal confidence coefficients. This function effectively solves the modal interference problem of the traditional weighted average method by dynamically evaluating the information entropy density of each modal feature and adaptively adjusting the fusion weights. The formula for the sound wave propagation path optimization equation is: In the formula: In the formula, di represents the Euclidean distance from the i-th probe to the defect point P, v is the sound velocity in the material, ti is the measured propagation time, and λ is the spatial continuity constraint coefficient. This equation introduces a spatial gradient regularization term on the basis of the traditional least squares method, which effectively suppresses the positioning jump caused by noise and realizes stable three-dimensional reconstruction of internal defects in complex structural components.

[0039] In step S5, the inspection report is automatically generated in PDF format, including defect type code, 3D coordinate positioning data, and calculated equivalent defect size, with a coordinate positioning accuracy of ±0.3mm. Data storage uses an industrial-grade encryption protocol, with a local storage period of no less than 30 days. If integrated with a cloud platform, the inspection data is uploaded in real time via HTTPS protocol, supporting SQL queries and trend analysis. The sorting system triggers pneumatic marking devices or robotic arms for sorting based on the defect level, with a marking error radius of less than 2mm and a sorting response time of no more than 0.5 seconds. Before being conveyed to the buffer area, qualified workpieces must undergo secondary photoelectric sensor verification to ensure a sorting accuracy rate higher than 99.8%.

[0040] From the above, we can conclude that: In this invention, by combining ultrasonic technology with optical imaging, the method can simultaneously detect surface and internal defects in a single operation. Ultrasonic waves penetrate the interior of the steel profile, accurately identifying hidden defects such as cracks and pores; the optical camera module captures visible problems such as surface scratches and corrosion in real time. This integrated detection mode avoids the cumbersome step-by-step process required by traditional methods. For example, it eliminates the need for first performing magnetic particle testing of surface defects and then using ultrasonic scanning of the internal structure, significantly shortening the detection cycle. It is particularly suitable for scenarios with high detection efficiency requirements, such as steel structure construction and bridge engineering.

[0041] In this invention, the fully automated design effectively reduces the impact of human intervention. The conveyor rollers are precisely controlled by servo motors to move the steel profiles at a uniform speed, with a speed deviation of less than ±0.1 m / min, ensuring the continuity of scanning in the detection area. Intelligent algorithms automatically analyze ultrasonic echo signals and optical image data, determining the defect level according to preset standards, avoiding subjective misjudgments from manual visual inspection. For example, for the identification of microcracks, the system automatically classifies them using feature thresholds, significantly improving the consistency of detection results when different operators use the same equipment.

[0042] In this invention, the modular design allows the equipment to flexibly adapt to steel profiles of different specifications. The spacing of the probe array can be adjusted manually or electrically via a mechanical structure, covering the inspection needs from narrow angle steel to wide H-beams; the anti-slip layer and drive parameters of the conveyor rollers can be dynamically optimized according to the weight and length of the workpiece. For example, when inspecting H-beams, adaptation can be completed simply by adjusting the probe layout and conveyor speed, without replacing core components, significantly reducing equipment modification costs and time.

[0043] In this invention, the non-contact inspection technology avoids direct contact with the steel profile surface throughout the entire process. The optical camera module uses long-distance, high-resolution imaging, eliminating the need for developing agents or coupling agents; the ultrasonic probe achieves non-contact scanning through air coupling or electromagnetic induction, completely solving the problems of magnetic suspension residue in traditional magnetic particle testing or chemical reagent contamination in penetrant testing. This method is particularly suitable for precision profiles with polished or coated surfaces, ensuring that the inspection process does not cause secondary damage to the workpiece.

[0044] In this invention, the integrated data processing system automatically converts inspection results into standardized reports, including the three-dimensional coordinates, equivalent dimensions, and type classification of defect locations, and supports local storage and cloud synchronization. Managers can quickly retrieve historical data using keyword searches, such as filtering records by time, material batch, or defect type, simplifying the quality traceability process. Furthermore, the cloud backup function facilitates cross-departmental collaborative analysis; for example, engineering teams can directly access inspection data to assess structural safety without relying on paper documents or manual summarization, significantly improving management efficiency.

[0045] In this invention, an ultrasonic probe array combined with a wavelet packet decomposition algorithm accurately captures the acoustic reflection characteristics of defects such as internal cracks and pores. An optical camera module uses a lightweight convolutional neural network to analyze surface scratches, corrosion, and other image features in real time, simultaneously detecting internal and external defects and avoiding the efficiency losses associated with step-by-step operations in traditional methods. A multimodal feature fusion algorithm is employed to dynamically adjust the weights of ultrasonic and optical data, combined with a cross-attention mechanism to eliminate the risk of misjudgment from single detection methods. For example, the correlation between surface scratches and internal cracks is automatically analyzed through a dual-channel neural network, improving the accuracy of defect classification. Adjustable probe spacing and a dynamic acoustic beam incident angle design, along with adaptive adjustment of conveyor roller parameters, enable rapid switching between I-beam and H-beam detection.

[0046] In this invention, the system automatically matches the probe layout and scanning path for different profile cross-sections, eliminating the need for manual recalibration. Non-contact optical imaging combined with air-coupled ultrasonic technology avoids the chemical contamination or ultrasonic coupling agent residue problems associated with traditional magnetic particle testing, making it particularly suitable for non-destructive testing of surface-coated or precision-machined profiles. Based on an improved particle swarm optimization algorithm, the 3D defect localization technology generates a digital report with coordinates, achieving a defect location error of less than 0.3 mm. Cloud synchronization supports real-time access across multiple terminals, simplifying the quality traceability process; for example, timestamps can be used to quickly locate the inspection records for a specific batch of profiles.

[0047] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the term "comprising" or any other variations thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0048] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A non-destructive testing device for steel profiles, characterized in that: The device includes: a detection device body (1), a conveying mechanism (2), a detection unit (3), and a control module (4); The main body of the detection device (1) is supported by a frame structure, which includes a base, columns and beams; the conveying mechanism (2) includes a conveying roller and a drive motor. The conveying roller is installed on the base, and the drive motor is connected to the conveying roller via a chain. The detection unit (3) includes an ultrasonic probe array and an optical camera module. The ultrasonic probe array is installed below the crossbeam via an adjustable bracket, and the optical camera module is fixed in the middle of the crossbeam. The control module (4) includes a data processing unit and a human-machine interface, which are installed in the control box on the side of the device.

2. The non-destructive testing method for steel profiles as described in claim 1, characterized in that: The ultrasonic probe array uses 8 wideband probes ranging from 0.5MHz to 10MHz.

3. The non-destructive testing method for steel profiles as described in claim 1, characterized in that: The optical camera module uses a 5-megapixel industrial camera.

4. The non-destructive testing method for steel profiles as described in claim 1, characterized in that: The conveyor rollers are made of 304 stainless steel and covered with a polyurethane anti-slip layer; the drive motor is a 1.5kW servo motor.

5. A non-destructive testing method for steel profiles, characterized in that: The non-destructive testing method includes the following steps: S1: Placement and positioning of steel profiles: Place the steel profile to be inspected at the beginning of the conveyor roller, ensuring that it is aligned with the roller, and fix its position with the anti-slip layer to prevent it from shifting during the inspection process; S2: Detection parameter settings: Input detection parameters through the human-computer interaction interface, including material type, detection speed, ultrasonic frequency range and imaging resolution of the optical camera module, and simultaneously select whether to enable the laser ranging module to monitor thickness changes; S3: Automated Synchronous Detection After the detection program is started, the conveyor rollers drive the steel profile into the detection area at a constant speed; the ultrasonic probe array scans internal defects through the principle of sound wave reflection, the optical camera module captures surface defects based on machine vision, and if the optimization scheme is enabled, the laser ranging module records the profile thickness data in real time. S4: Data Fusion and Defect Identification The control module receives ultrasonic signals and optical image data, and through algorithm fusion analysis, automatically identifies the defect type, location and size, determines whether it is qualified based on preset standards, and generates a defect map containing three-dimensional positioning information. S5: Report Output and Workpiece Sorting After the inspection is completed, the system automatically generates a digital inspection report and sends it to the local terminal or cloud platform through the control module; the inspected workpieces are output to the designated area by the conveyor rollers, and the unqualified products are automatically marked or sorted.

6. The non-destructive testing method for steel profiles as described in claim 5, characterized in that: In step S1, the steel profile must be placed horizontally at the starting end of the conveyor roller. The roller surface is made of 304 stainless steel and covered with a polyurethane anti-slip layer, with a friction coefficient of not less than 0.8, to ensure that there is no slippage or offset during workpiece transmission. The positioning accuracy requires that the longitudinal alignment error be less than ±1.5mm and the lateral offset not exceed 5% of the roller width. The drive motor is a 1.5kW servo motor, and the speed fluctuation rate is less than 0.5% through a closed-loop control system to ensure the stability of feeding in the subsequent detection area. The workpiece length is suitable for a range of 2m to 12m, and the maximum load capacity is 3 tons.

7. The non-destructive testing method for steel profiles as described in claim 5, characterized in that: In step S2, the material type needs to be selected from a pre-set database through the human-machine interface, covering 20 common steel grades such as Q235 and Q345. The system automatically loads the corresponding sound velocity and density physical parameters. The detection speed is adjustable from 0.5 to 2 m / min, with a recommended standard speed of 1 m / min to balance efficiency and accuracy. The ultrasonic probe array is configured with 8 independent channels, covering a frequency range of 0.5 MHz to 10 MHz. The low-frequency band is used to detect deep defects, while the high-frequency band captures near-surface microcracks. The optical camera module uses a 5-megapixel global shutter industrial camera with a frame rate matching the transmission speed to ensure that at least 10 effective pixels are collected per millimeter of workpiece surface. If the thickness monitoring function is enabled, the laser ranging module needs to be preheated and calibrated before the workpiece enters the detection area, and the measurement accuracy needs to reach ±0.05 mm.

8. The non-destructive testing method for steel profiles as described in claim 5, characterized in that: In step S3, the conveyor rollers drive the workpiece into the detection area at a constant linear speed, with the speed deviation controlled within ±0.1m / min; the ultrasonic probe array adopts a matrix layout, with an adjacent probe spacing of 50mm, and the sound beam incident angle is automatically adjusted according to the profile cross-sectional shape to ensure 100% sound wave coverage; the pulse repetition frequency of each probe is set to 5kHz, and the sampling depth is matched to more than twice the maximum thickness of the profile; the optical detection system uses coaxial light source illumination, with the light intensity set to be adjustable from 2000 to 5000 lux, and the exposure time is dynamically optimized according to the surface reflectivity to ensure that the image grayscale value distribution is within the effective range of 80-220; when thickness monitoring is enabled, the laser line scanning frequency must be synchronized with the conveying speed, and no less than 3 thickness data points are collected per millimeter of length.

9. The non-destructive testing method for steel profiles as described in claim 5, characterized in that: In step S4, the acoustic wave reflection signal acquired by the ultrasonic probe array is first analyzed in the time and frequency domain, and the defect echo features are extracted using an improved wavelet packet decomposition algorithm. A multi-modal feature joint discrimination model is constructed. The ultrasonic feature vector and the image feature vector are input into a dual-channel neural network, and feature interaction is achieved through a cross-attention mechanism. The contribution weight of each mode is dynamically adjusted using a gated fusion unit. For the generation of three-dimensional positioning information, a spatial positioning algorithm based on the acoustic wave propagation time difference is proposed. Through the collaborative measurement of multiple probe arrays, a set of geometric constraint equations for the acoustic wave propagation path is established, and the defect spatial coordinates are quickly solved using an improved particle swarm optimization algorithm. Finally, through the defect size quantization module, a composite calculation model of the defect volume is established by combining the acoustic wave attenuation coefficient and the image pixel resolution to achieve sub-millimeter-level precision defect quantification assessment. The formula for the multimodal feature fusion weight function is as follows: Where Fu and Fv represent ultrasonic and visual feature vectors, respectively, Wu and Wv are learnable projection matrices, and α and β are modal confidence coefficients; this function effectively solves the modal interference problem of the traditional weighted average method by dynamically evaluating the information entropy density of each modal feature and adaptively adjusting the fusion weights. The formula for the sound wave propagation path optimization equation is: In the formula: di represents the Euclidean distance from the i-th probe to the defect point P, v is the sound velocity of the material, ti is the measured propagation time, and λ is the spatial continuity constraint coefficient. This equation introduces a spatial gradient regularization term on the basis of the traditional least squares method, which effectively suppresses the positioning jump caused by noise and realizes stable three-dimensional reconstruction of internal defects of complex structural components.

10. The non-destructive testing method for steel profiles as described in claim 5, characterized in that: In step S5, the inspection report is automatically generated in PDF format, which includes defect type code, three-dimensional coordinate positioning data and equivalent defect size calculation value, with coordinate positioning accuracy reaching ±0.3mm.

Images