Optimization of measurement points for damage detection

Dynamical assessment is often employed to evaluate structures in civil engineering. This non-destructive surveillance technique became an attractive approach in order to show the existence of damages and to predict its location. The modal parameters of the structure are determined by modal identification methods without any prior knowledge of the nature of the excitation. The application of damage detection and localization techniques is performed by comparing the variation of the modal parameters between the states before and after damage. To do so, various methods can be used. In this paper, only the mode shape curvature and strain energy methods are considered because of the computation of the curvature in theirs algorithms. A discretization of the structure is needed on account of the location of the sensors and their number which generates a drawback. Commonly, the finest the discretization the better the approximate solution will be sufficiently accurate. However, several experimental analyses have shown this fact cannot be established in the damage detection domain. Indeed, this computation leads us to truncation errors including the effects of measurement noise, which leads to a loss of accuracy at the time of damage detection and localization. This paper aims to show the possibility to improve the accuracy of both proposed methods. The idea consists in minimizing the effect of noise and truncation errors by finding an optimal discretization interval. The values of this interval depend on several parameters, boundary conditions, bending mode, noise level. Numerical tests are carried out considering the presence of noise and damage simulations are performed on the tested beam. Furthermore, using the SysIdent-LCPC toolbox, several experimental tests are presented for the Merlebach’s beam in order to test the quality and the efficiency of the selected discretizations. The results are encouraging in order to go forwards the investigations in this direction