Damage Characterization by Model Updating Techniques Using a Derivative-Free Global Optimization Method

This work presents a damage characterization approach for composite structures, based on a global optimization technique, combining data retrieved from a finite element method, FEM, model and experimental signals from sensors (e.g., accelerometers). Damage diagnosis is based on vibrational methods. The proposed approach is able of simultaneously locating, determining the type of damage (changes on the stiffness, mass, damping) and outputting its severity (size and variation in material properties). The method demands automated pre/post-processing operations of dynamic FEM models, using Abaqus® scripting capabilities and continuous data exchange between Abaqus® and the external global optimization module based on Particle Swarm method. This is implemented into PRODDIA® – Structural Systems Health Management platform that manages the data and information flows. The method is firstly validated by a simulation approach and then an experimental case study is presented, validating its appropriateness and effectiveness to damage assessment on composite structures.