Analysis of the Magnetic Characteristics and Intelligent Identification Model of Defects in Small-Diameter Pipeline Elbow

Abstract

Abstract. Elbow plays a critical role in pipeline systems, and defects can lead to potential gas and oil leakage accidents. Magnetic flux leakage (MFL) detection is an efficient method for identifying pipeline defects. The unsaturated magnetic field in small-diameter pipe elbow was investigated. An image enhancement algorithm and intelligent identification method for MFL defects in small-diameter pipe elbows was proposed. The findings indicate that there is a distinct difference in magnetization intensity between straight and curved pipes, with uneven magnetic fields observed in curved pipes. Furthermore, defect size emerged as the primary factor affecting magnetic flux density, while defect location and the radius of curvature altered the distribution of magnetic flux density within the pipe. An image enhancement algorithm for defect leakage signals was proposed, resulting in a 6.73% increase in average detection accuracy after the incorporation of spatial pyramid dilated convolution (SPD-conv) and convolutional block attention modules. The test accuracies for general metal loss, pit metal loss, and circumferential groove were found to be 96.41%, 93.75%, and 94.08%, respectively, which is able to satisfy the demand for intelligent identification of leakage defects in small diameter pipe elbow.