Data-driven physical parameter estimation for lumped mass structures from a single point actuation test

Data-driven subspace system identification techniques developed from control theory provide a rich set of analytical tools to estimate state-space models of structural systems. Furthermore, a recent technological advance in physical parameter extraction directly from the identified state-space models paves a way for new developments in experimental structural dynamics. In this study, a practical issue of data-driven physical parameter estimation for lumped mass structures is discussed. To cope with difficulties related to dynamic tests of multiple actuations, this study proposes a signal processing strategy to estimate physical parameters from a dynamic test of single point actuation. To demonstrate the proposed strategy, an experimental study is conducted. A three-story shear frame structure is excited by a modal shaker with an input force measured and the corresponding displacement responses at each floor are scanned using a laser Doppler vibrometer (LDV). Based on the proposed strategy, physically interpretable mass and stiffness matrices, as well as non-proportional damping matrix are estimated solely from the measured experimental data.