Structural Parameter Identification Using Damped Transfer Matrix and State Vectors

A new method for identification of structural parameters is proposed using Damped Transfer Matrices (DTM) and state vectors. A new transfer matrix is derived for continuous mass systems including the damping parameters. The state vector at a location is the sum of the internal and external contributions of displacements, forces, and moments at that point, when it is multiplied with the transfer matrix, state vector at the adjacent location is obtained. The structural identification algorithm proposed here involves prediction of displacement responses at selected locations of the structure using DTM and compares them with the measured responses at the respective locations. The mean square deviations between the measured and predicted responses at all locations are minimized using a nonclassical optimization algorithm, and the optimization variables are the unknown stiffness and damping parameters in the DTM. A nonclassical heuristic Particle Swarm Optimization algorithm (PSO) is used, since it is especially suited for global search. This DTM algorithm with successive identification strategy is applied on one element or sub-structure of a structure at a time and identifies all the parameters of adjacent elements successively. The algorithm is applied on numerically simulated experiments of two structures such as ten degrees of freedom lumped mass system and a cantilever with seven finite elements and one sub-structure of a nine member frame structure. Also this algorithm is verified experimentally on a sub-structure of a fixed beam. The main advantage of this algorithm is that it can be used for the local identification in a zone in a structure without modeling the entire global structure.