Bayesian model identification of higher-order frequency response functions for structures assembled by bolted joints

Abstract

This paper proposes a procedure to identify a stochastic Bouc-Wen model for describing the dynamics of a structure assembled by bolted joints considering vibration data. The proposed identification approach is expressed into a Bayesian framework to take into account the data fluctuations related to uncertainties in the measurement process. The calibration of the model parameters uses the analytical expressions of the higher-order frequency response functions (FRFs) for approximating experimental measurements. The Metropolis-Hastings algorithm is employed for approximating posterior distributions. Once calibrated, the applicability of the probabilistic Bouc-Wen model is evaluated, and its dynamical behavior is compared with experimental measurements from the bolted structure. The results show that the stochastic version of the Bouc-Wen model can predict with adequate agreement, including hysteretic effects, the output of the jointed structure considering several excitation amplitudes.