An experimental verification of damage detection process using some of novel optimization techniques such as Latin hypercube sampling and swarm-based algorithms is presented. The algebraic differences between damage variables of numerical model and the test structures are formulated as an objective function which has to be minimized to identify damage location and its severity in the process of model updating. The profiles of modal frequency shifts become damage-sensitive features in conjunction with structural or damage variables such as mass or stiffness of numerical model. The iterative process which exploits the proposed population-based optimization algorithms successfully identifies local mass changes by updating damage variables to fit in modal data from test structures such as cantilevered beam and multi-bay truss frame.
展开▼
