LAUSR

Abstract The guided-wave-based non-destructive testing (NDT) technologies have been widely used for long range inspection of pipelines. Due to the dispersion and multi-modes characteristics of guided waves, the signal interpretation relies on the knowledge of dispersion curves. In present work, a field-measurement-based method is proposed for the inverse estimation of dispersion curves to achieve prior-knowledge-free guided wave inspections. Different from the commonly used 2D Fourier transform (2DFT) technologies, the proposed method applies bilevel optimization in the 3-D spectrum domain—frequency, axial wavenumber, and circumferential order—and remains valid in the presence of interference among multiple modes. The method is firstly verified by numerical simulation. The bias and uncertainty of the estimation are quantitatively evaluated by weighed mean bias (WMB) and weighted root mean square (WRMS), respectively. However, the experimental results exhibit distinct differences from the theoretically calculated dispersion curves. Compared to traditional technologies that utilized the theoretical dispersion curves, the proposed prior-knowledge-free technology is capable of delivering more decent results when applied for defect imaging. The effect of the implementation details of the field measurement, including the scanning density, the spacial coverage and the positioning errors, on the estimation accuracy are discussed.