Statistical model-based optimization for damage extent quantification

Damage localization and quantification constitute different aspects of structural damage diagnosis, which are particular interest in the Structural Health Monitoring field. Therein, a classical solution is model updating, where parameters finite element possibly damaged structure optimized to match with corresponding estimated from its vibration responses. To avoid ill-posedness updating problem, can be treated separately. First, information about regions or clusters elements obtained by method. Then, this used reduce number for quantification. A framework combining advantages methods optimization proposed paper. For exploration clustered physical space, stochastic algorithm coupled evaluation statistical properties MAC frequency differences between numerical modes an adequate treatment data-based uncertainties. Herein, development estimate important step, based on recent quadratic that adapted context inner product mode shape shape. This formulation objective function as well data-driven stopping criterion search. The validated simulations beam model, at multiple locations quantified up clustering precision.