A framework for backbone experimental tracking : Piezoelectric actuators, stop-sine signal and Kalman filtering

•The excitation is performed by piezoelectric patches.•The excitation signal is based on stop-sines measurement protocol (tuned on one mode).•The signal processing for parameter identification is based on extended Kalman filters.

This paper deals with the use of piezoelectric patches for nonlinear dynamic identification. The patches are glued on the structure to identify amplitude-dependent damping and natural frequency; their positions are defined in order to perform the excitation concentrated on the first bending mode. Their locations on the structure allow to perform “stop sines” tests, as, unlike electrodynamic shakers, piezos are embedded on structures and do not modify the studied structure after the excitation signal is switched off. Although, despite the piezo and the stop-sine, the signal is still modulated by other frequency components or polluted by random signals, a post processing with the extended Kalman Filter allows a very good determination of the modal damping and the natural frequency, especially when they depends on the free vibration amplitude.