Multiple stochastic Volterra models for damage detection
Résumé
An intrinsically nonlinear operation can complicate the damage detection process. Furthermore, the data variation related to actual measured signals makes the use of stochastic approaches essential, improving the reliability of the methods. Therefore, a stochastic version of the Volterra series based on the idea of multiple models (MM) is proposed to describe the behavior of the nonlinear structures in reference conditions, and statistical tools are implemented to introduce a hypothesis test to detect damage considering the presence of uncertainties.