Detection, localization, and quantification of crack-type damage in composites structures using vibration-based modal analysis

Issues such as energy security, deep decarbonization, and sustainable development have been a topic of world interest in recent years. Wind power is one of the main technologies enabling the transition to a cleaner and carbon-free energy sector. As wind power is deployed on a more massive scale, remote and automatic monitoring of wind turbines and their components becomes increasingly important. Currently, the blades of wind turbines are not routinely monitored. Implementing structural health monitoring techniques would allow knowing the health status of the structure while it is in operation. This would increase the reliability and lifetime of the wind turbine. In structural health monitoring vibration-based modal analysis methods are a powerful tool for detecting failures in a variety of structures. However, most of the studies so far have only focused on failure localization in simple systems such as beams or plates. In addition, they do not present information on the characteristics of the damages. This research has contributed to turning the detection, localization, and quantification of damage in composite structures into a robust technique that is ready for applications in a wide range of fields. First, a new method is presented based on modal analysis, wavelet denoising, and continuous wavelet transform in 2D. The method allows for locating and measuring the size characteristics of crack-type damage in thin-walled composite beams (TWCB), used as a proxy for wind turbine blades. This work significantly improved the required signal-to-noise ratio (at 15 dB) compared to similar works in the literature, also demonstrating the use of the technique to measure the length of cracks and their orientations, something that no other work had reported before. The experimental part was carried out by taking measurements in the TWCB employing a scanning laser vibrometer (SLV) on a bi-dimensional measurement grid. It was shown that the experimental results very closely matched the results obtained with finite-element modeling. This study shows the reliability of automatic fault detection using an array of sensors.