Quasi-Static Correction of Modally Reduced Order Models for System Inversion in Structural Dynamics

Abstract

A wide variety of system inversion algorithms has been proposed in the literature, tackling the problem of force identification, parameter estimation, and state/response estimation. Recently, much focus has gone to recursive joint input-state estimation and joint inputstate-parameter estimation, where the forces applied to the structure, the corresponding system states, and its parameters are simultaneously estimated in a recursive fashion. In order to reduce the computational load, these techniques are frequently used in combination with modally reduced order models. This paper shows that a model order reduction can lead to large estimation errors in the system inversion caused by disregarding the contribution of the so-called out-of-band modes. A recently developed computationally efficient quasi-static correction technique, which can be used in state-space modeling, is implemented and evaluated for both joint input-state estimation and joint input-state-parameter estimation. The evaluation is based on numerical simulations. It is shown that the quasi-static correction significantly reduces the estimation errors introduced by the model order reduction. 4386 COMPDYN 2019 7 th ECCOMAS Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering M. Papadrakakis, M. Fragiadakis (eds.