X-ray nondestructive testing method and system for composite insulator production process

Abstract

This invention discloses an X-ray non-destructive testing method and system for composite insulator production, belonging to the field of power equipment manufacturing quality inspection technology. The method includes the following steps: detecting the insulator's arrival signal and acquiring the current conveying speed of the production line using a photoelectric sensor; dynamically calculating the X-ray tube exposure time and X-ray incident angle based on the conveying speed; receiving the X-ray signal transmitted through the insulator and converting it into a digital X-ray image; an image processing module extracting the detection target area of ​​the sheath-core rod bonding interface layer from the digital X-ray image and performing image quality judgment on the detection target area; comparing the image feature parameters with a pre-calibrated defect judgment benchmark, and determining whether the interface bonding quality is qualified based on the comparison result. This method enables direct, rapid, and non-destructive online detection of the composite insulator interface bonding quality on the production line, thereby detecting internal bubbles, debonding, and partial discharge traces in advance and preventing potential failures.

Description

Technical Field

[0001] This invention relates to an X-ray non-destructive testing method and system for composite insulator production, belonging to the field of power equipment manufacturing quality inspection technology. Background Technology

[0002] Composite insulators mainly consist of three parts: a glass fiber reinforced epoxy resin core rod, a silicone rubber sheath and sheds, and end fittings. Due to their advantages such as light weight, pollution resistance, high strength, and ease of maintenance, they are widely used in power systems. However, brittle fracture (referred to as "brittle fracture") of composite insulators during operation is one of the most serious safety hazards. Numerous accident analyses indicate that the root cause of brittle fracture lies in poor adhesion between the sheath and the core rod interface: uneven application of the interface adhesive, insufficient curing, or improper surface treatment of the core rod during production leads to the presence of micropores and debonding areas at the interface. During long-term operation, these micro-defects, under the influence of a strong electric field, induce partial discharge, gradually eroding the material and forming carbonization channels, ultimately leading to the loss of mechanical strength of the core rod.

[0003] Currently, the main methods for quality testing of composite insulators include the following.

[0004] Visual inspection can only detect surface defects and cannot detect internal interface problems. Ultrasonic testing requires a coupling agent, is slow, has low efficiency for continuous testing on production lines, and is insufficiently sensitive to interface debonding. Power frequency withstand voltage testing can only indirectly reflect insulation strength and cannot pinpoint the specific location and type of defects. Infrared thermometry can only detect surface temperature rise and is not sensitive to early interface defects.

[0005] For example, patent application publication number CN115825112A discloses a method and system for detecting internal defects in composite insulators. This scheme is based on microwave detection technology, deriving a wave impedance model by establishing Maxwell's equations, and using the reflectivity equation of microwaves at the interface between the silicone rubber layer and the epoxy resin layer to determine the defect situation. However, this scheme has the following shortcomings: First, the spatial resolution of microwave detection is limited by the microwave wavelength (centimeter level), and its ability to identify micropore defects with a diameter of less than 1 mm is limited; second, this method only uses a single parameter of reflectivity for judgment, and cannot distinguish between different types of interface defects such as micropores and debonding; third, this scheme is an offline detection method and does not involve an online continuous detection mechanism linked to the production line speed.

[0006] Patent application publication number CN107478728A discloses a non-destructive testing method for composite insulators. This method employs an ultrasonic testing system that uses an ultrasonic transducer to emit ultrasonic signals that are incident on the composite insulator under test. The interface adhesion is analyzed based on the resonant frequency of the ultrasonic echo signal. However, this method requires the application of an ultrasonic coupling agent, resulting in low testing efficiency. Furthermore, the ultrasonic signal attenuates significantly in silicone rubber materials, leading to insufficient sensitivity for detecting deep interface defects.

[0007] In summary, existing technologies lack a method for directly, quickly, and non-destructively inspecting the bonding quality of composite insulator interfaces online on a production line, especially a dedicated technical solution that can be linked to the production line speed, dynamically match inspection parameters, and quantitatively determine the type and size of defects through grayscale statistical analysis. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings of the prior art and provide an X-ray non-destructive testing method and system for composite insulator production, thereby detecting internal bubbles, debonding and partial discharge traces in advance and preventing potential failures.

[0009] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the X-ray non-destructive testing method for the composite insulator production process of the present invention includes the following steps: S1: The composite insulator is conveyed to the X-ray inspection station along the production line. The insulator arrival signal is detected by the photoelectric sensor and the current conveying speed v of the production line is obtained. S2: The synchronous control module dynamically calculates the X-ray tube exposure time t based on the current conveying speed v of the production line. e And select the incident angle θ of the X-ray so that the motion blur of the sheath-core interface area on the imaging surface does not exceed the preset pixel threshold, and control the X-ray source to emit penetrating X-ray beams within the matched time window; S3: Receives X-ray signals transmitted through the insulators via a flat panel detector located on the other side of the production line and converts them into digital X-ray images; S4: The image processing module extracts the detection target area of ​​the sheath-core bonding interface layer from the digital X-ray image, performs image quality judgment on the detection target area, and calculates the image feature parameters of the area when the image quality meets the preset standard. S5: Compare the image feature parameters with the pre-calibrated defect judgment benchmark, and determine whether the interface bonding quality is qualified based on the comparison result, and identify the defect type; S6: Feedback the judgment result to the production line control system to drive the sorting mechanism to remove or mark non-conforming products.

[0010] Furthermore, in step S2, the exposure time t e The matching relationship with the transmission speed v satisfies: Where d is the thickness of the silicone rubber sheath, and k is the fuzziness tolerance coefficient; the incident angle θ of the X-ray is in the range of 15°≤θ≤35°, so that the X-ray beam penetrates the sheath-core bar interface layer at an inclined angle, and the thickness magnification effect of the interface layer in the inclined projection direction is used to improve the imaging contrast of defects; the synchronous control module communicates with the main control system of the production line through industrial Ethernet, obtains the production line speed information in real time, and dynamically adjusts the tube voltage U and tube current I parameters of the X-ray source so that the penetration ability of the X-ray is matched with the material thickness and density distribution of the insulator.

[0011] Furthermore, the detection target area is a strip-shaped region continuously cut along the insulator axis with the outer edge of the core rod and the inner edge of the sheath as the boundary. The width of this region is no more than one-third of the core rod diameter. N frames of images are continuously acquired at the same detection position, where N≥3.

[0012] Furthermore, in step S4, the image quality determination includes the following sub-steps: S41: Calculate the overall signal-to-noise ratio of the digital X-ray image. When the signal-to-noise ratio is lower than a preset threshold, it is determined to be an invalid image and a resampling is triggered. S42: Detect the grayscale histogram distribution pattern of the detection target area. When the grayscale value is concentrated in the range of less than 20% of the full scale or more than 80% of the full scale, it is determined that the exposure parameters are abnormal and the X-ray source parameters are adaptively adjusted. S43: Verify the spatial position of the detection target area using a template matching algorithm. When the position deviation exceeds the preset allowable value, trigger the synchronization control module to correct the detection timing.

[0013] Furthermore, the image feature parameters include: the average gray value of pixels within the detected target area. Gray-scale variance Statistical mean of gray-scale gradient modulus and the average gray value of the reference area of ​​the mandrel substrate. The substrate reference area is selected from a uniform region inside the core rod far from the interface layer in the same frame image of the self-detection target area; the grayscale gradient mode The calculation method is as follows: calculate the horizontal gradient for each pixel within the detection target area. and vertical gradient The gradient magnitude is obtained. Then, take the statistical mean of the gradient magnitudes of all pixels.

[0014] Furthermore, in step S5, the defect judgment criterion is pre-calibrated using grayscale statistical samples of intact bonding interfaces of the same specification, including the grayscale mean. Gray standard deviation Upper limit of grayscale variance of intact interface The defect determination includes the following steps: S51: Calculate the normalized grayscale deviation ; S52: When and At that time, it was determined that the interface had microporous defects, among which, Determine the negative threshold for stomata; S53: When and When the interface is determined to have a debonding air gap defect, then... The positive threshold for determining debonding; S54: When and When the interface bonding is deemed satisfactory; S55: When At that time, it was determined that the interface had carbonization traces defects, among which The preset outlier variance threshold is used, and .

[0015] Furthermore, the method further includes the following steps: S7: Implement a denser detection strategy for key detection areas, where the key detection areas are the total length of the insulator extending downwards along the axial direction from the high-voltage end of the sheath. proportionality coefficient The section at which the length is doubled, i.e., the key detection area, is [length missing]. The voltage borne by this section during operation is no less than the proportion of the voltage borne by the insulator as a whole. The encryption detection strategy involves increasing the X-ray exposure frame rate for the region at the same transmission speed, thereby increasing the number of imaging frames for that region. No less than the number of frames in other regions of times, that is ,in ; S8: The original X-ray images, defect judgment results, defect types, defect location coordinates and product identification information generated during the inspection process are automatically archived in the database to generate a traceable quality inspection report with a unique traceability code; S9: Non-conforming products rejected by the sorting mechanism are processed according to a preset ratio. Perform destructive cutting verification, compare the measured defect status of the cutting verification with the automatic judgment result of step S5, and calculate the false detection rate. and false positive rate ;when or At that time, the average gray value in the defect judgment criterion is automatically adjusted. Gray standard deviation and outlier variance threshold This allows the detection system's judgment criteria to adaptively adapt to the current production process batch, among which... and These are the preset thresholds for false negative and false positive rates.

[0016] Furthermore, the method also includes automatically retrieving the corresponding detection parameter configuration table by reading the product model identifier when switching insulator product models. The detection parameter configuration table includes the tube voltage range. Tube current range Exposure time limit , range of incident angles Detection target area ratio coefficient Grayscale determination threshold , and This enables rapid replacement testing between insulators of different specifications.

[0017] Secondly, the present invention provides an X-ray non-destructive testing system for a composite insulator production line that implements the X-ray non-destructive testing method in the composite insulator production process, comprising: An X-ray emitting module for generating a penetrating X-ray beam includes an X-ray source, a collimator, a shield, and a high-voltage controller; the collimator shapes the beam to a size that matches the cross-section of the insulator. An imaging receiving module is used to receive X-ray signals transmitted through an insulator and convert them into digital images. It includes an amorphous silicon flat panel detector, a filter, a collimating grating, a signal amplifier, and an analog-to-digital converter. The filter covers the surface of the flat panel detector to filter low-energy scattered rays. The collimating grating is located in front of the flat panel detector and its opening size matches the size of the detection target area to shield scattered X-rays from non-detection areas. The synchronization control module is used to coordinate the timing of X-ray emission and image acquisition and link it with the production line speed. It includes a PLC controller, photoelectric sensors and a timing generator. The image processing module is used to perform image quality assessment, target area extraction, image feature parameter calculation, and defect classification, including a GPU-accelerated computing unit; The communication module is used to enable data transmission between the various modules; The safety protection module is used to ensure the radiation safety of operators.

[0018] Furthermore, the communication module includes a dual-link redundancy structure consisting of a first radio frequency communication link and a second radio frequency communication link for real-time transmission of high-definition X-ray images, and a backup link for communication protection when the main link is subject to electromagnetic interference; the safety protection module includes a shielding structure, a radiation dose monitor arranged inside and outside the shielding structure, a protective door interlocking device interlocked with the X-ray source, an emergency stop button located inside and outside the shielding structure, and a radiation warning light.

[0019] Compared with existing technologies, the X-ray non-destructive testing method and system for composite insulator production of the present invention exhibits the following beneficial effects in terms of technical performance and practical application: (1) Dynamic matching of motion parameters to eliminate motion blur: By dynamically matching the step exposure time and the transmission speed and the tilting projection design of the X-ray incident angle, the problem of imaging blur caused by product movement in X-ray online inspection is solved.

[0020] (2) Defect classification and quantification are achieved by statistical judgment of grayscale dual threshold: By using the two-dimensional joint judgment criteria of normalized grayscale deviation and grayscale variance, three different types of interface defects, namely micropores, debonding gaps and carbonization traces, can be distinguished at the same time. Compared with the existing technology that judges the existence of defects by only the single parameter of microwave reflectivity, the judgment method of the present invention has higher defect classification ability and recognition accuracy.

[0021] (3) Closed-loop feedback calibration achieves adaptive adjustment: By comparing the results of destructive cutting verification with automatic judgment, the gray mean, gray standard deviation and abnormal variance threshold are dynamically adjusted, so that the detection system can adapt to the differences between batches of different production processes, which solves the technical problem that the judgment standard is fixed and cannot adapt to the fluctuation of production process in the existing technology.

[0022] (4) Enhancing the detection rate of key parts by intensifying detection in key areas: By implementing an intensified detection strategy on a quarter section of the high-voltage side (with an imaging frame count no less than M times that of other areas), the defect detection rate of the key section with the highest voltage is specifically improved. Compared with the uniform detection method of all insulators in the existing technology, the detection accuracy of key areas is significantly improved under the same detection time.

[0023] (5) Image quality three-level judgment ensures data reliability: Through the three-level quality judgment mechanism of signal-to-noise ratio check, gray-scale histogram distribution check, and template matching position verification, invalid images caused by abnormal exposure parameters, product shaking and other factors are effectively filtered out, ensuring the reliability of image data entering the defect judgment stage and reducing the false detection rate.

[0024] (6) Online full inspection and quick changeover: The system can achieve 100% online full inspection in sync with the production line speed, and can realize quick changeover inspection between insulators of different specifications through the detection parameter configuration table. Compared with the offline detection method, it greatly improves the detection efficiency and production line adaptability.

[0025] (7) Multi-parameter image feature enhancement for robustness of identification: By extracting multi-dimensional image feature parameters such as average gray value, gray variance, and gray gradient modulus, it has stronger defect identification robustness compared with single gray feature, reducing the risk of misjudgment caused by X-ray source fluctuations or batch differences in product materials. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating the overall process of the method in this embodiment of the invention. Figure 2 This is a schematic diagram of the composite insulator structure and X-ray detection system in an embodiment of the present invention; Figure 3 This is a block diagram of the X-ray inspection system for the composite insulator production line in an embodiment of the present invention; In the diagram: 1. Upper fitting; 2. Core rod; 3. Umbrella skirt; 4. Sheath; 5. Lower fitting. Detailed Implementation

[0027] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0028] Example 1: like Figure 2 As shown, the composite insulator mainly consists of an upper fitting 1, a core rod 2, sheds 3, a sheath 4, and a lower fitting 5. The core rod 2, made of glass fiber reinforced epoxy resin through a pultrusion process, bears the mechanical load. The sheath 4 and sheds 3 are made of silicone rubber material, forming a continuous insulating layer on the surface of the core rod 2 through an injection molding process. The upper fitting 1 and lower fitting 5 are fixed to both ends of the core rod 2 by compression fitting. The adhesive interface layer between the sheath 4 and the core rod 2 is the core target area for testing.

[0029] After X-rays penetrate the insulator, the differences in absorption of the rays by different materials create a grayscale contrast image on the detector. The linear attenuation coefficient of core rod 2 (epoxy resin + glass fiber) is approximately 0.28 cm. -1 @120kV, the line attenuation coefficient of sheath 4 (silicone rubber) is approximately 0.19cm. -1@120kV. Because the core material has a higher density and effective atomic number than the sheath material, it has a stronger ability to absorb X-rays. Therefore, in X-ray transmission images, the core region appears as a relatively dark gray, while the sheath region appears as a relatively bright gray. A normally bonded interface layer exhibits a continuous and uniform transition zone from the sheath gray to the core gray.

[0030] When defects exist in the interface layer, the X-ray absorption characteristics of the defective region change: the microporous defects are filled with air, and the linear attenuation coefficient of air is only about 0.02 cm. -1 @120kV, far lower than normal silicone rubber and epoxy resin, therefore the radiation is absorbed and reduced at the pore locations, increasing the intensity of the radiation received by the detector, which appears as local bright spots (higher grayscale values) in the image; the debonding air gap also appears as local bright spots due to air filling, but since the debonding gap is usually flat and extends along the interface layer, its grayscale abnormal area appears as a narrow and elongated shape; the X-ray absorption characteristics of carbon (atomic number 6) in the carbonization traces are significantly different from those of silicon (atomic number 14) in the silicone rubber matrix, and the absorption of radiation in the carbonization area is uneven, appearing as a textured area with chaotic grayscale values ​​in the image.

[0031] Taking a 220kV composite insulator production line as an example, the production line cycle time is 5 seconds between each product, the conveying speed is v=0.2m / s, and the specific testing process is as follows: like Figure 1 As shown, the X-ray non-destructive testing method for composite insulators during the production process of this invention includes the following steps: S1: The composite insulator is conveyed to the X-ray inspection station along the production line. The insulator arrival signal is detected by the photoelectric sensor and the current conveying speed v of the production line is obtained. In this embodiment, the insulator moves to the detection station with the conveyor belt at a speed of 0.2m / s. The photoelectric sensor (with a response time better than 0.5ms) detects that the front end of the insulator is in place, triggers the positioning signal, and records the current conveyor speed v=0.2m / s.

[0032] S2: The synchronous control module dynamically calculates the X-ray tube exposure time t based on the transmission speed v. e And select the incident angle θ of the X-ray so that the motion blur of the sheath-core interface area on the imaging surface does not exceed the preset pixel threshold, and control the X-ray source to emit penetrating X-ray beams within the matched time window; In this embodiment, the synchronization control module adjusts according to the sheath thickness. Fuzzy tolerance coefficient Calculate the upper limit of exposure time Actual settings At this point, the motion blur is controlled to be within 1 pixel. Select the ray incident angle. By using oblique projection, the projected thickness of the interface layer on the imaging surface was magnified from 12 mm to approximately 13.2 mm, improving the visibility of the interface layer. The X-ray source tube voltage was adjusted to 120 kV and the tube current to 3 mA. After confirming that the protective door was closed and the dose monitoring was normal, a trigger signal was sent to the X-ray source and detector.

[0033] S3: Receives X-ray signals transmitted through the insulators via a flat panel detector located on the other side of the production line and converts them into digital X-ray images; In this embodiment, the X-ray source emits a beam that penetrates the insulator with an exposure time of 20ms. The beam passes sequentially through the sheath 4, the interface layer, and the core rod 2. The detector simultaneously acquires the transmission image and generates a 16-bit grayscale digital image with a size of 1024×3072 pixels and a single frame data size of approximately 6MB.

[0034] S4: The image processing module extracts the detection target area of ​​the sheath-core bonding interface layer from the digital X-ray image, performs image quality judgment on the detection target area, and calculates the image feature parameters of the area when the image quality meets the preset standard. In this embodiment, the original image is transmitted to the image processing unit via a 5.8GHz radio frequency communication link. A three-level image quality assessment is then performed. S41: Signal-to-noise ratio check: Calculate the overall signal-to-noise ratio of the image. Higher than the preset threshold The image is deemed valid. S42: Histogram check: The grayscale histogram distribution range of the detected target area is between 28% and 72% of the full scale, and no exposure parameter adjustment has been triggered; S43: Position verification: Template matching deviation is 0.8 pixels, which is lower than the preset tolerance value. Pixel and location verification passed.

[0035] After the image quality is deemed acceptable, the feature parameters of the target area image are calculated. The measured values ​​for this frame are shown in Table 1.

[0036] Table 1 Measured values ​​of image feature parameters

[0037] S5: Compare the image feature parameters with the pre-calibrated defect judgment benchmark, and determine whether the interface bonding quality is qualified based on the comparison result, and identify the defect type; In this embodiment, the defect judgment benchmark was obtained by pre-calibrating 50 sets of samples of intact bonding interfaces of the same specification. The calibration results are shown in Table 2.

[0038] Table 2 Defect Judgment Criteria Calibration Results

[0039] Calculate the normalized gray deviation . Since and , the interface bonding is determined to be qualified.

[0040] S6: Feed the determination result back to the production line control system, and drive the sorting mechanism to remove unqualified products or mark unqualified products.

[0041] To verify the accuracy of the defect determination algorithm of the present invention, composite insulators containing typical interface defects are detected. The physical mechanism of X-ray imaging is as follows: when there are micro air pores in the interface layer, the linear attenuation coefficient of air in the pores (0.02 cm -1 ) is far lower than that of normal silicone rubber (0.19 cm -1 ) and epoxy resin (0.28 cm -1 ), less radiation is absorbed when the radiation passes through the air pores, the intensity of radiation received by the detector increases locally, and a bright spot region with high gray value is formed on the image, therefore, the average gray value G I is significantly higher than the mean value μ of the normal interface, and the gray deviation ΔG presents a large negative value (due to the high gray value of the mandrel reference area, the gray change of the micro air pore area relative to the intact interface after normalization presents a negative deviation). When there is a debonding gap in the interface layer, the local radiation absorption is also reduced due to air filling, a bright spot is presented on the image with a high gray value, therefore, ΔG presents a large positive value. When there are carbonization traces on the interface layer, the X-ray absorption characteristics of carbon element in the carbonized area are significantly different from that of the silicone rubber matrix, and the internal absorption of the carbonized area is uneven, resulting in sharp fluctuations of gray value in space, and the gray variance increases significantly. Based on the above physical mechanism, adopting ΔG and two-dimensional joint determination criterion can effectively distinguish three different types of defects.

[0042] Table 3 Comparison table of various defect determination results and cutting verification

[0043] It can be seen from Table 3 that the method of the present invention can recognize micro air pore defects ( and ), debonding gap defects ( and ), carbonization trace defects ( ) and qualified interfaces ( and ) four states can all be accurately identified, and all are consistent with the actual situation after destructive cutting verification.

[0044] Based on the principle of high-voltage electric field distribution, the section of a composite insulator extending downwards along the axis from the high-voltage end of the sheath to one-quarter of the total length of the insulator bears more than 75% of the voltage borne by the entire insulator, making it the most sensitive inducing area for interface bonding defects. Therefore, an intensive detection strategy is implemented for this key area.

[0045] The proportionality coefficient β = 0.25, the voltage withstand ratio γ = 75%, and the encryption frame rate multiplier M = 2 were set. For a 220kV composite insulator with a length of 2m, the key detection area was the section 0.5m downwards from the high-voltage end. Table 4 shows a comparison of the detection parameters at the same transmission speed v = 0.2m / s.

[0046] Table 4 Comparison of detection parameters between key and non-key areas

[0047] As shown in Table 4, the axial sampling density in key areas is twice that in non-key areas, and the axial size of the smallest detectable defect is reduced by half, which significantly improves the detection accuracy of key parts.

[0048] Non-conforming products rejected by the sorting mechanism are processed according to a preset ratio. Destructive cutting verification was performed. For example, the test data for a specific batch of products is shown in Table 5.

[0049] Table 5 Comparison of detection performance before and after closed-loop feedback calibration.

[0050] As shown in Table 5, the false negative rate before calibration Exceeding the preset threshold The system will automatically The values ​​were adjusted from 128 to 126 and s from 8.5 to 8.0, narrowing the pass / fail judgment range. After calibration, the false negative rate dropped to 3.5% and the false positive rate dropped to 18%, both within the preset threshold range.

[0051] When the production line switches from 220kV insulators to 110kV insulators, the system automatically retrieves the corresponding testing parameter configuration table by reading the product model identifier via RFID. The testing parameter configurations for different insulator specifications are shown in Table 6.

[0052] Table 6. Configuration of Testing Parameters for Insulators of Different Specifications

[0053] The system completes the switching and loading of all parameters within 2 seconds, without the need for manual reset, enabling rapid model change detection.

[0054] Example 2: Based on Example 1, such as Figure 3As shown, the X-ray non-destructive testing system for composite insulator production lines of this invention consists of six core modules. The signal flow between the modules is as follows: the synchronization control module acts as the central nervous system of the system, detecting the product position through photoelectric sensors and precisely controlling the X-ray emission timing of the X-ray emission module and the image acquisition timing of the imaging receiving module; the X-ray beam generated by the X-ray emission module is collimated and then irradiates the composite insulators on the production line; the imaging receiving module is located on the other side of the production line, receiving the X-ray signals that pass through the insulators and converting them into digital image data; the wireless communication module transmits the data to the image processing module in real time; the image processing module performs image quality judgment, target area extraction, feature parameter calculation, and defect classification judgment; the judgment results are fed back to the production line control system; and the safety protection module provides fully enclosed radiation protection for the inspection station.

[0055] The image processing module, based on a GPU-accelerated computing platform, executes the following processing flow: (a) Preprocessing stage: 3×3 median filtering is used to remove impulse noise, and adaptive histogram equalization (with contrast enhancement limited and CLIP parameter set to 2.0) is used to enhance the overall contrast of the image, thereby enhancing the grayscale gradient of the interface layer by about 1.8 times.

[0056] (b) Target Area Localization Stage: A template is established based on prior insulator geometric parameters. A normalized cross-correlation matching algorithm is used to locate the boundary positions of the outer edge of the core rod and the inner edge of the sheath in the entire image. The template size is 64×64 pixels, the matching threshold is set to 0.85, and the localization accuracy reaches the sub-pixel level. Based on this, a strip-shaped target area containing the adhesive interface layer is accurately extracted. The width of the target area is set to 1 / 3 of the core rod diameter (approximately 8 mm), and the axial length covers the entire length of the insulator.

[0057] (c) Feature extraction stage: Calculate the average gray value of pixels within the detection target area. Gray-scale variance Statistical mean of gray-scale gradient modulus and the average gray value of the reference area of ​​the mandrel substrate. The substrate reference area is selected from a uniform region inside the core rod at a distance of more than 5 mm from the interface layer in the same frame image of the self-detection target area, with a size of 20×20 pixels.

[0058] (d) Defect Judgment Stage: According to the defect judgment criteria in Example 1, based on the normalized grayscale deviation... and gray variance Based on the relationship, the four states of micropores, debonding gaps, carbonization traces and qualified interfaces are identified according to the judgment rules in Example 1, and the judgment results are output in the format of Table 7.

[0059] Table 7 Defect Judgment Output Data Format

[0060] Communication module: A 5.8GHz frequency band radio frequency communication is used as the primary link, with a theoretical transmission rate of no less than 100Mbps and a measured single-frame image transmission time of approximately 0.5ms. A LoRa long-range communication module (frequency 470MHz, transmission rate 50kbps) is configured as a backup link to ensure low-rate backhaul of critical judgment results when the primary link is subject to electromagnetic interference. All communication data is encrypted using the AES-256 algorithm, with an encryption latency of no more than 0.1ms.

[0061] Safety Protection Module: The entire testing station is housed within a lead-lined room with a shielding efficiency of no less than 99%. One radiation dose monitor (range 0.01 μSv / h to 10 mSv / h, accuracy ±10%) is installed both inside and outside the lead-lined room to display the ambient radiation level in real time. An alarm is triggered and the X-ray source is automatically shut off when the ambient radiation dose exceeds 1 μSv / h. The lead-lined room's protective door is equipped with an electromagnetic lock linked to the X-ray source—the X-ray source cannot start if the door is not closed, and automatically shuts off the X-ray source within 20ms if the door is accidentally opened. An emergency stop button is located both inside and outside the lead-lined room; pressing it will shut off the X-ray source within 5ms.

[0062] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A non-destructive X-ray testing method for composite insulators during production, characterized in that, The steps include the following: S1: The composite insulator is conveyed to the X-ray inspection station along the production line. The insulator arrival signal is detected by the photoelectric sensor and the current conveying speed v of the production line is obtained. S2: The synchronous control module dynamically calculates the X-ray tube exposure time t based on the current conveying speed v of the production line. e And select the incident angle θ of the X-ray so that the motion blur of the sheath-core interface area on the imaging surface does not exceed the preset pixel threshold, and control the X-ray source to emit penetrating X-ray beams within the matched time window; S3: Receives X-ray signals transmitted through the insulators via a flat panel detector located on the other side of the production line and converts them into digital X-ray images; S4: The image processing module extracts the detection target area of ​​the sheath-core bonding interface layer from the digital X-ray image, performs image quality judgment on the detection target area, and calculates the image feature parameters of the area when the image quality meets the preset standard. S5: Compare the image feature parameters with the pre-calibrated defect judgment benchmark, and determine whether the interface bonding quality is qualified based on the comparison result, and identify the defect type; S6: Feedback the judgment result to the production line control system to drive the sorting mechanism to remove or mark non-conforming products.

2. The X-ray non-destructive testing method for composite insulator production process according to claim 1, characterized in that, In step S2, the exposure time t e The matching relationship with the transmission speed v satisfies: Where d is the thickness of the silicone rubber sheath, and k is the fuzziness tolerance coefficient; the incident angle θ of the X-ray is in the range of 15°≤θ≤35°, so that the X-ray beam penetrates the sheath-core bar interface layer at an inclined angle, and the thickness magnification effect of the interface layer in the inclined projection direction is used to improve the imaging contrast of defects; the synchronous control module communicates with the main control system of the production line through industrial Ethernet, obtains the production line speed information in real time, and dynamically adjusts the tube voltage U and tube current I parameters of the X-ray source so that the penetration ability of the X-ray is matched with the material thickness and density distribution of the insulator.

3. The X-ray non-destructive testing method for composite insulator production process according to claim 1, characterized in that: The detection target area is a strip-shaped region continuously cut along the insulator axis with the outer edge of the core rod and the inner edge of the sheath as the boundary. The width of this area is no more than one-third of the core rod diameter. N frames of images are continuously acquired at the same detection position, where N≥3.

4. The X-ray non-destructive testing method for composite insulator production process according to claim 1, characterized in that: In step S4, the image quality determination includes the following sub-steps: S41: Calculate the overall signal-to-noise ratio of the digital X-ray image. When the signal-to-noise ratio is lower than a preset threshold, it is determined to be an invalid image and a resampling is triggered. S42: Detect the grayscale histogram distribution pattern of the detection target area. When the grayscale value is concentrated in the range of less than 20% of the full scale or more than 80% of the full scale, it is determined that the exposure parameters are abnormal and the X-ray source parameters are adaptively adjusted. S43: Verify the spatial position of the detection target area using a template matching algorithm. When the position deviation exceeds the preset allowable value, trigger the synchronization control module to correct the detection timing.

5. The X-ray non-destructive testing method for composite insulator production process according to claim 4, characterized in that: The image feature parameters include: the average gray value of pixels within the target area. Gray-scale variance Statistical mean of gray-scale gradient modulus and the average gray value of the reference area of ​​the mandrel substrate. The reference region for the core rod substrate is selected as a uniform region inside the core rod far from the interface layer in the same frame image of the self-detection target area; the statistical mean of the grayscale gradient modulus. The calculation method is as follows: calculate the horizontal gradient for each pixel within the detection target area. and vertical gradient The gradient magnitude is obtained. Then, take the statistical mean of the gradient magnitudes of all pixels.

6. The X-ray non-destructive testing method for composite insulator production process according to claim 1, characterized in that: In step S5, the defect judgment criterion is pre-calibrated using grayscale statistical samples of intact bonding interfaces of the same specification, including the grayscale mean. Gray standard deviation Upper limit of grayscale variance of intact interface ; The defect determination includes the following steps: S51: Calculate the normalized grayscale deviation ; S52: When and At that time, it was determined that the interface had microporous defects, among which, Determine the negative threshold for stomata; S53: When and When the interface is determined to have a debonding air gap defect, then... The positive threshold for determining debonding; S54: When and When the interface bonding is deemed satisfactory; S55: When At that time, it was determined that the interface had carbonization traces defects, among which The preset outlier variance threshold is used, and .

7. The X-ray non-destructive testing method for composite insulator production process according to claim 6, characterized in that: It also includes the following steps: S7: Implement a denser detection strategy for key detection areas. The key detection area is the section extending downwards along the axial direction from the high-voltage end of the sheath to a point where the length of the insulator is proportional to a factor β (L times the total length L). The denser detection strategy involves increasing the X-ray exposure frame rate for this area at the same transmission speed, thus increasing the number of imaging frames for this area. No less than the number of frames in other regions M times, that is ,in ; S8: The original X-ray images, defect judgment results, defect types, defect location coordinates and product identification information generated during the inspection process are automatically archived in the database to generate a traceable quality inspection report with a unique traceability code; S9: Non-conforming products rejected by the sorting mechanism are processed according to a preset ratio. Perform destructive cutting verification, compare the measured defect status of the cutting verification with the automatic judgment result of step S5, and calculate the false detection rate. and false positive rate ;when or At that time, the average gray value in the defect judgment criterion is automatically adjusted. Gray standard deviation and outlier variance threshold This allows the detection system's judgment criteria to adaptively adapt to the current production process batch, among which... and These are the preset thresholds for false negative and false positive rates.

8. The X-ray non-destructive testing method for composite insulator production process according to claim 6, characterized in that: The method also includes automatically retrieving the corresponding detection parameter configuration table by reading the product model identifier when switching insulator product models. The detection parameter configuration table includes the tube voltage range. Tube current range Exposure time limit , range of incident angles Detection target area ratio coefficient Grayscale determination threshold , and This enables rapid replacement testing between insulators of different specifications.

9. An X-ray non-destructive testing system for a composite insulator production line implementing the X-ray non-destructive testing method in the composite insulator production process according to any one of claims 1 to 8, characterized in that, include: X-ray emission module, used to generate penetrating X-ray beams, includes an X-ray source, collimator, shield, and high-voltage controller; The collimator shapes the ray beam to a size that matches the cross-section of the insulator; An imaging receiving module is used to receive X-ray signals transmitted through an insulator and convert them into digital images. It includes an amorphous silicon flat panel detector, a filter, a collimating grating, a signal amplifier, and an analog-to-digital converter. The filter covers the surface of the flat panel detector to filter low-energy scattered rays. The collimating grating is located in front of the flat panel detector and its opening size matches the size of the detection target area to shield scattered X-rays from non-detection areas. The synchronization control module is used to coordinate the timing of X-ray emission and image acquisition and link it with the production line speed. It includes a PLC controller, photoelectric sensors and a timing generator. The image processing module is used to perform image quality assessment, target area extraction, image feature parameter calculation, and defect classification, including a GPU-accelerated computing unit; The communication module is used to enable data transmission between the various modules; The safety protection module is used to ensure the radiation safety of operators.

10. The X-ray non-destructive testing system for composite insulator production lines according to claim 9, characterized in that, The communication module includes a dual-link redundancy structure consisting of a first radio frequency communication link and a second radio frequency communication link, used for real-time transmission of high-definition X-ray images, and a backup link for communication protection when the main link is subject to electromagnetic interference; the safety protection module includes a shielding structure, a radiation dose monitor arranged inside and outside the shielding structure, a protective door interlocking device interlocked with the X-ray source, an emergency stop button set inside and outside the shielding structure, and a radiation warning light.

Images