A robust and computationally efficient guided wave tomography of pipes

Guided wave ultrasonic tomography of pipe usually uses the helical modes propagated between parallel transducer arrays to compensate for the missing incident angle . Nevertheless, separating mixed signals from various helical modes during inversion is challenging, and this method necessitates multiple replications of the two-dimensional unfolding area , leading to high computational costs. It also faces practical problems such as environmental noise, transducer differences, and poor coupling . To address these problems, this paper proposes a finite difference method that utilizes helical mode information without multiple replications or the need to separate signals. Additionally, an objective function based on zero-delay cross-correlation is employed to mitigate the effects of noise, transducer inconsistencies, and coupling issues that distort signal amplitudes. The proposed method reduces the inversion space of high-order helical modes of the pipeline to the area of a two-dimensional unfolding, and improves the reliability and accuracy of inversion through objective function optimisation. In numerical simulations, the algorithm achieves thickness loss imaging with errors less than 0.1 mm and demonstrates good performance in experimental tests .