| An important objective of structural health monitoring systems is to identify the state of the structure and to detect the damage when it occurs. Analysis techniques for damage identification of structures, based on vibration data measured from sensors, have received considerable attention. The traditional time-domain analysis techniques require that all the external excitations and accelerations respondes be available, which may not be the case for most structural health monitoring systems, due to practical limitations. The adaptive sequential non-linear least-square estimate method with unknown (unmeasured) excitations (inputs) and unknown (unmeasured) acceleration responses (outputs) has been proposed to identify the parameters of the systems and the variety by Yang et al (2007). The necessary condition of this method is the number of DOFs must be greater than the total number of unknown inputs and unknown outputs. Though, the sufficient condition, that how many known inputs and known outputs be needed to identify the time varying parameters of the system of all DOFs, has not been proposed. Therefore, firstly, the sufficient condition is theoretically obtained by analyzing the rank of data matrix which related to structural parameters. Then, numerical simulations with multiple degrees of freedom (MDOF) linear structures are performed to identify the parameters of the structures. The number of unknown inputs and unknown outputs is increased step by step, and the identification results of the simulations are the same as the theoretical ones. And on-line experiment based on 3-DOF shear-beam model is conducted to validate the correctness of the theoretical results. Finally, based on the static condensation method, the 12-DOF finite element model is condensed into a 6-DOF system, and the rotational DOFs have been condensed, and the equation turn into normal equation, to verity the theoretical sufficient analytical method mentioned the paper. |
PDF Full Text Request