Nonlinear System Identification of a Multi-story Building with Geometrical Nonlinearity Using a Deterministic Output-Only-Data Approach

Nonlinear system identification based on output-only data is challeng-ing since the stochastic approaches require the structure to be excited by random input with a uniform Gaussian distribution. This paper applies a deterministic output-only approach to parameter estimation of a linear multi-story specimen with an amplitude-dependent geometrical nonlinearity. The approach is inde-pendent of the input type, value, and number but requires the excitation to be applied away from the nonlinearity. The vibration responses to high-amplitude excitations are taken into a subspace-based identification algorithm that simulta-neously yields both nonlinear and underlying linear parameters. The process is verified by comparing the underlying linear parameters with the linear modal pa-rameters of the structure under low-amplitude excitation. The results indicate a superior accuracy of the estimated parameters in the simulation and an acceptable confidence range for the experimental test.