Multi–perturbed analytical models for updating and damage detection

The two problems of analytical model updating and damage detection in structures are identical in both formulation and limitations. One of the main limitations in this type of problem is the lack of sufficient information (equations) and the tendency of the mathematical models to be ill–conditioned. Multiple tests under different mass, stiffness or boundary conditions to correct an analytical model have previously been proposed to increase the amount of useful information and to improve the condition of equations. Such an approach can be costly, time consuming and prone to experimental errors. A novel approach based on single–test and multiple analytical models is presented here. The analytical model of the structure can serve as the basis for updating, in addition to several pseudo–analytical models. These pseudo–analytical models are created by randomly perturbing the original analytical model in order to maximize information and numerical stability. Examples are included in support of the proposed technique.