Rasgado-Moreno Carlos-Omar, Rist Marek, Land Raul, Ratassepp Madis
Department of Civil Engineering and Architecture, Tallinn University of Technology, Ehitajate tee 5, 19086, Tallinn, Estonia.
Thomas Johann Seebeck Department of Electronics, Tallinn University of Technology, Ehitajate tee 5, 19086, Tallinn, Estonia.
Ultrasonics. 2025 Apr;148:107560. doi: 10.1016/j.ultras.2024.107560. Epub 2025 Jan 1.
Pipe bends are recognized as critical areas susceptible to wall thinning, a phenomenon instigated by abrupt changes in the fluid flow direction and velocity. Conventional monitoring techniques for bends typically depend on localized ultrasonic measurements of thickness. While these methods are effective, they can be time-consuming compared to the use of permanently installed transducers, a strategy employed in guided wave tomography (GWT). GWT provides the advantage of identifying and quantifying damage within a specified area by processing waves that are both generated and received by a set of transducers. In this study, we implement a GWT method based on full waveform inversion (FWI) for a high-resolution thickness reconstruction of a steel pipe bend. The wavefield in the bend section, made artificially anisotropic, is modeled using Thomsen parameters in the two-dimensional domain. This enhances its integration with the FWI algorithm. A numerical investigation was conducted to evaluate the efficacy of FWI in the presence of a defect as a function of its circumferential position. Additionally, an experimental evaluation was performed to reconstruct a defect artificially created on a pipe bend with a diameter (d) of 220 mm and a bend radius of 1.5d, and a defect with a diameter of 100 mm and a depth of 47%. The results indicate that the FWI method can effectively reconstruct the thickness map of smooth defects, regardless of their location, and it is particularly effective for defects situated closer to the extrados position.