Time Domain Identification of Structural Parameters and Input Time History Using A Substructural Approach

This paper presents a substructural approach to parameter identification of structures, with the simultaneous identification of structural parameters and input time history of the applied excitation. The substructural approach reduces the computational effort when dealing with large structures. The substructural parameters, including the unknown interface forces at the ends of the substructure, are identified iteratively. The locations of input forces must be known; however, their magnitudes can be unknown. The method requires the measurement of accelerations at the interior DOFs but not interface DOFs of the substructure. Numerical simulations are performed for three examples, viz. global models of a 15 DOF shear building model, a planar truss of 55 members and a cantilever beam of 20 elements. The examples are excited with harmonic, random and impulse excitations. The effect of noisy data is studied. Even with noise, the proposed substructural method is found to identify the structural parameters with appreciable accuracy and with a considerable saving of CPU time. However, the identification of damping parameters is found to be prone to more errors than for the stiffness parameters.