A Computer Vision-Based Approach for Non-contact Modal Analysis and Finite Element Model Updating

Computer vision-based techniques for modal analysis and system identification are rapidly becoming of great interest for both academic research and engineering practice in structural engineering. For instance, this is particularly relevant in fields such as bridge or tall building monitoring, where the large size of the structure would require an expensive sensor network, and for the characterisation of very slender, highly-flexible structural components, where physically-attached sensors cannot be deployed without altering the mass and stiffness of the system under investigation. This study concerns the latter case. Here, an algorithm for the full-field, non-contact extraction and processing of useful information from vibrational data is applied. Firstly, video acquisition is used to capture rapidly very spatially- and temporally-dense information regarding the vibrational behaviour of a high-aspect-ratio (HAR) prototype wing, with high image quality and high frame rate. Video processing is then applied to extract displacement time histories from the collected data; in turn, these are used to perform Modal Analysis (MA) and Finite Element Model Updating (FEMU). Results are benchmarked against the ones obtained from a single-point laser Doppler vibrometer (LDV). The study is performed on the beam-like spar of the wing prototype with and without the sensors attached to appreciate the disruptive effects of sensor loading. Promising results were achieved.