Abstract: Aiming at the problem of structural parameter identification and unknown input inversion under earthquake, this paper proposed a method of structural parameter identification and seismic input inversion using only partial absolute acceleration measurements as the observation. Firstly, based on transform space method and least squares method, structural parameters above the first story were identified directly. Then, the stiffness of the first story was identified based on the identified modal information. Finally, the dynamic  equation was transformed into the transform space in which the correlation model of seismic input at adjacent sampling time was established and a simplified input inversion algorithm is presented. Results of numerical simulation and shaking table tests on a two-story reinforced concrete frame show that the proposed method has high accuracy and robustness for linear systems, and for nonlinear systems, the inversion results reflect the low-frequency characteristics of seismic input which can be used on seismic performance deductive analysis for regional structures and buildings.

Key words: structural parameter identification, seismic input inversion, spectral characteristic, transform space method, least square method