Waveform Inversion of Ultrasonic data – first synthetic and laboratory experiments

Ultrasonic imaging methods are widely used in non-destructive testing (NDT) and structural health monitoring (SHM) to ensure the safety of critical infrastructure. This work focuses on a methodology that combines Full Waveform Inversion (FWI) and Reverse Time Migration (RTM) techniques to improve the imaging of ultrasonic data and analysis of complex concrete structures. This approach allows a deeper understanding of the internal properties and heterogeneities within complex materials. The European project I am involved in, "USES of novel UltraSonic and Seismic Embedded Sensors for the non-destructive evaluation and structural health monitoring of infrastructure and human-built objects" (USES2) (www.uses2.eu), aims to integrate novel sensor technologies, advanced processing and innovative imaging to improve the monitoring of industrially relevant applications in sectors such as energy, mobility and construction.

Ultrasonic testing provides critical insight into the interior of concrete structures, including thickness measurement, geometry determination, localization of embedded components, and material quality assurance. However, the heterogeneous and multiphase nature of concrete poses challenges such as wave dispersion, scattering and attenuation, which make it difficult to detect small defects and internal features. FWI and RTM overcome these limitations by using the full ultrasonic wave field, unlike conventional approaches that use only a portion of the measured data. As a result, these techniques enable high-resolution recovery of material properties and detailed imaging of complex internal structures.

The first results of this work are presented here. Data sets were acquired using the A1040 MIRA 3D Pro ultrasound tomography system on polyamide and concrete specimens with drilled holes. The recorded ultrasonic wavefields were analyzed and compared with the modeling results. To ensure consistency between simulated and recorded wavefields, the source signature was recovered using multiple approaches and the dominant signature for FWI was identified. FWI-based finite-difference modeling was then applied to the ultrasound data to estimate elastic properties such as shear wave velocity and density distributions. This analysis provided valuable information for identifying internal inhomogeneities and served as an initial model for RTM.

The next phase of this research will extend the methodology to ultrasonic data collected from more complex samples and in-situ structures. Significant effort will be focused on refining the imaging resolution of ultrasonic data using RTM. The RTM algorithm has the potential to improve resolution in imaging concrete media, particularly in resolving steeply dipping interfaces and complex structures.

In summary, this methodology, which integrates advanced FWI and RTM approaches to ultrasonic data, offers new possibilities for the diagnosis and detection of heterogeneities and defects in concrete structures. It provides high-resolution and consistent images for NDT and SHM applications, addressing critical challenges in these fields.