TIME-VARYING MODAL IDENTIFICATION BY MONTE CARLO FILTER

Abstract eng:
In this paper, changing dynamic characteristics of time-varying systems, which are modeled in the form of modal decomposition, are identified using Monte Carlo filters. Extending the ordinary modal decomposition form to the time-varying case, a time-varying MDOF system is represented as a superposition of time-varying SDOF systems, which have time-varying mode shape vectors, time-varying natural frequencies and time-varying damping ratios. The system equations and the non-Gaussian evolution models of modal parameters are merged into an augmented state space model. The resultant state space model is nonlinear and non-Gaussian, so that the state vector needs to be identified using nonlinear non-Gaussian estimators. A Monte Carlo filter is applied, which is one of the most promising approach to the nonlinear non-Gaussian state estimation problems. Numerical simulations for a simple 2-DOF system show that the proposed identification scheme can track the abrupt change of the modal characteristics even in noisy circumstances.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket