Characterization of Clustered Cracks at Weld Root Based on Uniform-Magnetic-Field Distortion Measurement

Girth welds in pipelines are critical links on which the pipeline system depends. After long-term operation, defects may appear at the weld root, which seriously threatens the safety of the system. Cracks are the most dangerous defect, especially multiple clustered cracks. Some closely spaced cracks are often misjudged as corrosion defects, leading to an error estimation of residual strength of pipelines and causing sudden accidents. The accurate characterization of closely spaced cracks is thus of great significance for subsequent hazard assessment. However, traditional nondestructive testing methods all have limitations because of the complex stress distribution and excess weld penetration at the weld root. To solve this problem, we propose to detect cracks using uniform magnetic field distortion (UMFD) method, which is highly sensitive to surface defects but insensitive to local variations in stress. This article proposes an analytical model of an isolated crack whose variables are merely the position and width of the crack. Furthermore, we provide a characterization algorithm based on this analytical model for evaluating the position and width of subcracks. Both the simulation and experimental results indicate that the proposed method accurately characterizes cracks. To assess the robustness of the proposed algorithm, additive white Gaussian noise with signal-to-noise ratios (SNRs) of 5 and 10 dB is superposed onto the raw data from the simulation. The results show that this method is robust against background noise.