Autocalibration method for guided wave tomography with undersampled data

This paper presents a baseline-free quantitative method for the imaging of corrosion flaws present in thin plates with under-sampled data. This method is based on the Hybrid Algorithm for Robust Breast Ultrasound Tomography (HARBUT) which is itself inherently baseline-free. However, in order to ensure that the incident field component becomes negligible in the reconstruction, a calibration step is necessary. Indeed, it is essential to rescale the data with respect to the acoustic model whether it be simulation data or experimental data. This calibration is usually performed by manually choosing a ray for which the domain of propagation is assumed sound. This can be problematic because this method is not automatic. Moreover, if the chosen ray happens to pass through a flaw, the resulting image will be of poor quality. This paper proposes an autocalibration method for the rescaling step. The field of application is Structural Health Monitoring (SHM) of critical structures with heavy constraints on both sensor intrusiveness and diagnostic reliability. In order to limit intrusiveness, a sub-sampled array of embedded guided waves sensors within the structure is used. Extensions to HARBUT are introduced to compensate for the aliasing caused by the undersampling. The benefits of these extensions are then assessed with numerical simulations and experimental datasets measured by a PZT network.