A large-scale defect eddy current testing method for station process pipeline

Abstract

The application discloses a station yard process pipeline defect large-lift magnetic flux leakage detection method and relates to the technical field of oil and gas pipeline safety. The steps are as follows: characteristic values are extracted by collecting pipeline magnetic flux leakage signals, mobile filtering, peak-valley extraction, width threshold filtering and sliding deduplication; real defects are identified by using the mean value and the standard deviation of multiple sensors; the safety of defects is graded and evaluated according to industry standards; based on the magnetic anomaly length, the defect depth and the lift height, the real size of the defect is inversed through a polynomial fitting model, so that the accurate positioning, classification and quantitative detection of pipeline defects under the conditions of large lift and with a coating layer are realized. The application provides a detection scheme for the operating state of a steel pipeline and has important significance for ensuring the safe and normal operation of the pipeline.

Description

Technical Field

[0001] This invention relates to the field of oil and gas pipeline safety technology, and in particular to a method for detecting magnetic flux leakage in process pipeline defects at stations. Technical Background

[0002] Steel pipelines in oil and gas stations are typically fitted with anti-corrosion layers, insulation layers, and outer protective structures to ensure operational safety and withstand the effects of the external environment. As service life increases, defects on the inner wall of the pipeline are mostly caused by the corrosive components and impurities in the transported medium, leading to corrosion and wear defects over long-term operation. Defects on the outer wall are mainly caused by electrochemical volumetric corrosion resulting from the failure of the coating layer.

[0003] Magnetic flux leakage (MF) detection methods are effective at detecting defects on pipe surfaces. Existing research largely focuses on wall-mounted detection conditions, primarily analyzing MF signals under strong magnetic saturation magnetization for defect inversion. While some researchers have optimized MF detection techniques to address lift-off interference, current improvements are mostly focused on single-stage optimizations and lack a systematic detection and inversion framework. It is necessary to develop a MF signal processing and defect size quantitative inversion method suitable for large lift-off conditions. This method should overcome the limitations of traditional strong magnetic saturation wall-mounted detection, enabling accurate inversion of corrosion defects on the inner and outer walls of pipes under unsaturated magnetization and multi-interference coupling conditions. Ultimately, this will achieve efficient external detection without interrupting production or peeling off the coating. Summary of the Invention

[0004] This invention proposes a quantitative inversion method for non-saturated magnetic flux leakage signals applicable to large-scale lifting conditions of pipelines without removing the cladding. By collecting basic information of the cladding pipeline and acquiring magnetic flux leakage signals under large-scale lifting conditions, the defect size is quantitatively inverted, thus constructing a method for processing magnetic flux leakage signals and quantitatively inverting defect size applicable to large-scale lifting conditions.

[0005] The technical solution provided by this invention to solve the above-mentioned technical problems is a method for detecting large-scale magnetic flux leakage defects in process pipelines in a station, comprising the following steps: S1. Collect basic information on the coated pipeline and conduct on-site inspections on the coated pipeline. The magnetic sensor array contains eight parallel magnetic sensors, arranged one every 15° along the circumference of the pipeline, to collect multi-directional magnetic field data from the magnetic sensor array. B ; S2. Apply moving filter to obtain the filtered first... i indivual( i =1, 2, 3, n) magnetic sensors in j Magnetic field in directions (including x, y, and z directions) The length of the sliding window is determined based on the dimensions of the defect magnetic anomaly. L The first data point in the sliding window corresponds to the mileage.s Extract data from the sliding window Find the magnetic field data within the sliding window. Each peak ( t Represents the peak value. k =1, 2, 3, n) and their mileage Then, using the same method, the magnetic signal is negativeized to find local maxima, and then negativeized again to obtain the valley values ​​of the magnetic field data within the window. ( b (representing the valley value) and its mileage ; S3, will z Window data of direction and The corresponding peak or valley values ​​are subtracted in ascending order of mileage, and the absolute value is taken to obtain the peak and valley values. (m=1, 2, 3, n) and the corresponding peak and valley widths If the peak-valley width Signals less than 60mm (preset interference threshold) are considered interference signals and will not be saved. The calculation formula is as follows: (1); S4 x Peak and valley width in the direction and z Peak and valley width in the direction Consistency, find and Within range x Valley with the smallest direction Then we get x Peak and valley values ​​in the direction The calculation formula is as follows: (2); save x Directional peak and valley values , x Directional peak and valley width , z Directional peak and valley values , z Directional peak and valley width and z Directional peak position ; S5. Move the first data in the window. s Increase L / 3. Reload data To enable window sliding, re-locate the windows within each window according to S4. x direction and zThe eigenvalues ​​of the direction; if obtained from two window extractions z If the mileage difference corresponding to the directional peak is less than 20 sampling points (corresponding to a length of 10mm), it is judged as a duplicate feature and discarded without being stored. S6. Based on the fact that the eight magnetic sensors arranged side by side are in the same position x Directional peak and valley values Sure x Orientation average peak-to-valley value As shown in equation (4). A standard deviation threshold is set. Compare the standard deviations corresponding to each local maximum point. and standard deviation threshold This allows us to determine whether the location is due to interference or a pipeline defect. When standard deviation Greater than the standard deviation threshold If the location is not obstructed, it is considered an interference; otherwise, it is considered a pipeline defect. (3); (4); S7. According to SY / T6151-2022 "Evaluation Method for Metal Loss Defects in Steel Pipelines," defects are irregular and require measurement of their depth using a pit depth gauge or ultrasonic thickness gauge to determine the deepest point of the defect. When the corrosion defect depth exceeds 80% of the pipe wall thickness, an immediate response is required; however, a response is not necessary when the defect depth is less than 10% of the pipe wall thickness. In reality, most corrosion defects fall between 10% and 80% of the pipe wall thickness. Therefore, it is necessary to calculate the maximum safe operating pressure and the maximum permissible operating pressure (MAOP) and compare their magnitudes. If the maximum safe operating pressure is less than the MAOP, an immediate response is required even if the depth is less than 80%. Establish a defect classification method for pipes with cladding layers based on eigenvalues: The feature value with the largest amplitude of each magnetic sensor is selected, and a trained classifier is used to determine the category of pipeline defect, locating all Class II and Class III defects. The maximum safe operating pressure and MAOP are calculated, and it is determined whether the Class II defect is acceptable. For some severe defects, their true length is calculated. S8. Both cylindrical and cubic defects in the pipeline have rectangular cross-sections along the axial direction. As shown in equation (5), based on half the length of the obtained magnetic anomaly. x (Acquired via magnetic sensor), defect depth h 1 (Output from the defect classification model) and lift-off height y (This can be obtained from the design data of the coating thickness) Half the length of the defect can be calculated. b Model parameters to be determined a , c ,d , e , f , g By collecting basic parameter information of the coated pipeline and simultaneously acquiring on-site measured data of the pipeline, the relevant undetermined parameters are determined by using a polynomial fitting method. (5); In the formula: b The length is half the defect length, m; h 1 Let m be the defect depth. y The lift-off height is measured in meters (m). x It is half the length of the magnetic anomaly, in meters (m). a , c , d , e , f , g These are the undetermined parameters for the model. Beneficial effects: This invention is particularly suitable for the rapid and accurate detection of corrosion defects in steel pipelines within the station without removing the coating layer. It can achieve detection under conditions of large-scale lifting without production interruption, effectively reducing construction and shutdown costs. Relying on multi-dimensional magnetic field feature filtering and array sensor feature discrimination mechanism, it accurately quantifies defect size and determines the hazard level, fully complying with relevant specifications for oil and gas pipeline inspection and evaluation, with outstanding detection reliability and field applicability. Attached Figure Description

[0006] Figure 1 Flowchart of a method for detecting magnetic flux leakage in process pipelines at a station.

[0007] Figure 2 This is a schematic diagram of on-site testing in one embodiment. Detailed Implementation

[0008] The present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. S10. Collect basic information on the cladding pipe: pipe diameter 273mm, wall thickness 8mm, pipe material L245, lifting height 40mm. S20. During installation, adjust the sensor array spacing, installing one sensor every 15°. The handheld device should be rolled axially along the specific clock direction of the pipeline and inspected at a speed not exceeding 5cm / s. Multiple axial inspections should be performed on key pipelines to cover all circumferential positions, thus forming a complete magnetic detection grid for pipeline defects. For pipelines with ordinary defect coatings, at least one axial inspection should be performed along the 6 o'clock direction. If magnetic anomalies caused by defects are found, the section should be marked as a key inspection area and subjected to full circumferential coverage inspection. S30. Export the collected magnetic field data in each direction. Calculate the collected data using equations (1) and (2). x Directional peak and valley values , x Directional peak and valley width , z Directional peak and valley values ; (1); (2); S40. Eliminate interference signals along the pipeline and extract the peak and valley values ​​of the pipeline defects in the x-direction using a defect discrimination method based on eigenvalues. =0.81Gs and peak-to-valley width =106.9mm and peak-valley values ​​in the z-direction =1.1Gs; S50, Based on known peak and valley values ​​in the x-direction =0.81Gs and peak-to-valley width =106.9mm and peak-valley values ​​in the z-direction =1.1Gs, and by substituting the pre-trained classification algorithm into the calculation, the classification result shows that the depth level of the defect is Level II; (3); (4); S60. Remove the coating layer from unacceptable Class II defects and all Class III defects. Use visual inspection combined with ultrasonic thickness measurement to locate pipe defects and determine the remaining wall thickness. Verify the accuracy of the magnetic flux leakage test results under the coating layer and confirm the defect depth. h 1 =2.7mm; S70, Based on half of the obtained magnetic anomaly length x =53.45mm, defect depth h 1 =2.7mm, lifting height y =40mm and the undetermined parameters of the fitted model a =4、 c =5、d =8、 e =4、 f =4, g=1. Half the length of the defect can be calculated using equation (5). b =100mm, which fits the actual situation; (5); S80. The above description is merely a technical solution of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made based on the technical essence of the present invention without departing from the scope of the technical solution of the present invention are included within the protection scope of the present invention.

Claims

1. A method for detecting large defects in process pipelines in a station, characterized in that, Includes the following steps: S1. Collect basic information on the coated pipeline and conduct on-site inspections on the coated pipeline. The magnetic sensor array contains eight parallel magnetic sensors, arranged one every 15° along the circumference of the pipeline, to collect multi-directional magnetic field data from the magnetic sensor array. B ; S2. Apply moving filter to obtain the filtered first... i indivual( i =1, 2, 3, n) magnetic sensors in j Magnetic field in directions (including x, y, and z directions) . The length of the sliding window is determined based on the size of the defect magnetic anomaly. L The first data point in the sliding window corresponds to the mileage. s Extract data from the sliding window Find the magnetic field data within the sliding window. Each peak ( t Represents the peak value. k =1, 2, 3, n) and their mileage Then, using the same method, the magnetic signal is negativeized to find local maxima, and then negativeized again to obtain the valley values ​​of the magnetic field data within the window. ( b (representing the valley value) and its mileage ; S3, will z Window data of direction and The corresponding peak or valley values ​​are subtracted in ascending order of mileage, and the absolute value is taken to obtain the peak and valley values. (m=1, 2, 3, n) and the corresponding peak and valley widths If the peak-valley width Signals less than 60mm (preset interference threshold) are considered interference signals and will not be saved. The calculation formula is as follows: （1）； S4 x Peak and valley width in the direction and z Peak and valley width in the direction Consistency, find and Within range x Valley with the smallest direction Then we get x Peak and valley values ​​in the direction The calculation formula is as follows: （2）； save x Directional peak and valley values , x Directional peak and valley width , z Directional peak and valley values , z Directional peak and valley width and z Directional peak position ; S5. Move the first data in the window. s Increase L / 3. Reload data To enable window sliding, re-locate the windows within each window according to S4. x direction and z The eigenvalues ​​of the direction; if obtained from two window extractions z If the mileage difference corresponding to the directional peak is less than 20 sampling points (corresponding to a length of 10mm), it is judged as a duplicate feature and discarded without being stored. S6. Based on the fact that the eight magnetic sensors arranged side by side are in the same position x Directional peak and valley values Sure x Orientation average peak-to-valley value As shown in equation (4). A standard deviation threshold is set. Compare the standard deviations corresponding to each local maximum point. and standard deviation threshold This allows us to determine whether the location is due to interference or a pipeline defect. When standard deviation Greater than the standard deviation threshold If the location is not obstructed, it is considered an interference; otherwise, it is considered a pipeline defect. （3）； （4）； S7. According to SY / T6151-2022 "Evaluation Method for Metal Loss Defects in Steel Pipelines," defects are irregular and require measurement of their depth using a pit depth gauge or ultrasonic thickness gauge to determine the deepest point of the defect. When the corrosion defect depth exceeds 80% of the pipe wall thickness, an immediate response is required; however, a response is not necessary when the defect depth is less than 10% of the pipe wall thickness. In reality, most corrosion defects fall between 10% and 80% of the pipe wall thickness. Therefore, it is necessary to calculate the maximum safe operating pressure and the maximum permissible operating pressure (MAOP) and compare their magnitudes. If the maximum safe operating pressure is less than the MAOP, an immediate response is required even if the depth is less than 80%. Establish a defect classification method for pipes with cladding layers based on eigenvalues: The feature value with the largest amplitude of each magnetic sensor is selected, and a trained classifier is used to determine the category of pipeline defect, locating all Class II and Class III defects. The maximum safe operating pressure and MAOP are calculated, and it is determined whether the Class II defect is acceptable. For some severe defects, their true length is calculated. S8. Both cylindrical and cubic defects in the pipeline are rectangular in cross-section along the axial direction. As shown in equation (5), based on half of the obtained magnetic anomaly length x (Acquired via magnetic sensor), defect depth h 1 (Output from the defect classification model) and lift-off height y (This can be obtained from the design data of the coating thickness) Half the length of the defect can be calculated. b Model parameters to be determined a , c , d , e , f , g By collecting basic parameter information of the coated pipeline and simultaneously acquiring on-site measured data of the pipeline, the relevant undetermined parameters are determined by using a polynomial fitting method. （5）； In the formula: b The length is half the defect length, m; h 1 Let m be the defect depth. y The lift-off height is measured in meters (m). x It is half the length of the magnetic anomaly, in meters (m). a , c , d , e , f , g These are the undetermined parameters for the model.

Images