REGULARIZED SOLUTION FOR STRUCTURE HEALTH MONITORING

Resumo

Structural damage is a key concern for several areas in civil engineering sector (buildings,
bridges), mechanical engineering, aerospace sector (aircrafts, rockets, satellites), automotive sector, and
many other Therefore, the structure health monitoring is an important topic research and operation. The
structural damage identification can be addressed as an inverse vibration problem. This inverse problem
can be formulated as a generalized least square problem, where the stiffness matrix must be identified in

order to have a best matching between measurements and the mathematical model added to a regulariza-
tion operator. We solve the forward problem using finite element method, and an entropic regularization

is also applied to the cost function. The optimization problem is solved by using a hybrid method, com-
bining a stochastic metaheuristic with a local searching method. The Multi-Particle Collision Algorithm

(MPCA) is the metaheuristic technique and the Hooke-Jeeves (HJ) direct search method complete the
hybrid optimizer. The proposed methodology is applied to different study cases showing good results. A
cantilever beam is the testing structure for the developed approach.