Time domain system identification using advanced iterative least Squares with unknown input method

Abstract

This paper introduces a new system identification method to estimate the structural parameters for nondestructive damage evaluation of structures. It is the Advanced Iterative Least-Squares Method, which identifies the structural parameters without kind of information on input excitation. It is a time-domain linear system identification method capable of identifying structures at the element level. Dynamic response needs to be measured for input. Initial structural parameters are also need to be assumed in advance with simple methods, from experience of similar structures or investigation of appearance. Input forces can be evaluated with the governing equation based on dynamic response and initial structural parameters. They make constraints of input excitations. The proposed method is verified with two examples, shear-type structures subject to various input forces including seismic load. Simulation result demonstrates that the advanced iterative least square with unknown input method can invariably identify the structural parameters with high precision. Copyright © (2006) by Earthquake Engineering Research Institute.