Damage detection based on a stochastic version of the Volterra series

Luis Gustavo Giacon Villani (1) , Samuel da Silva (1) , Americo Cunha Jr (2)

1 UNESP - Universidade Estadual Paulista Júlio de Mesquita Filho = São Paulo State University
2 UERJ - Universidade do Estado do Rio de Janeiro [Brasil] = Rio de Janeiro State University [Brazil] = Université d'État de Rio de Janeiro [Brésil]

This paper aims the damage detection in a clamped-free beam, presenting nonlinear behavior for large displacements, considering the data variation related with the presence of uncertainties. For this, a stochastic version of the Volterra series, expanded through a random version of the Kautz functions, is implemented, using Monte Carlo simulations to estimate the underlying statistics. The measurements were performed in different days and conditions, to consider the data variation and the presence of nonlinear behavior in the analysis. Two different damage indexes are proposed, based on the advantage of the use of Volterra series, related with the capability to separate linear and nonlinear contributions in the total system response. The results have shown that the nonlinear approach is more sensible to the presence of inserted damage (loss of mass), presenting better results than the classical linear approach.

Domaines: Vibrations [physics.class-ph]; Probabilités [math.PR]; Statistiques [math.ST]; Mécanique des structures [physics.class-ph]; Mécanique des solides [physics.class-ph]; Génie mécanique [physics.class-ph]

Luis Gustavo Giacon Villani, Samuel da Silva, Americo Cunha Jr. Damage detection based on a stochastic version of the Volterra series. 2018. ⟨hal-01811702⟩