Phd of KORDYLAS Layla

Abstract :

The project focuses on stochastics methods designed in order to identify main parameters and their uncertainties for two kinds of dynamics :
+ the dynamics of local joints (in terms of nonlinear stiffness and energy losses)
+ the operational modal analysis (OMA) under weak nonlinearity hypothesis
All the local joints are known to be subject to great nonlinearities but the dynamics of whole structures composed with these joints highlights weak nonlinearities. The weak nonlinearity in modal analysis means that modal damping ratios and natural frequencies are nonlinear regarding the excitation amplitude and environment conditions but that the mode shapes do not depend on the excitation amplitude.
Models and assumptions will be analyzed in the data assimilation process. We will try to check the intrinsic assumption of Gaussian processes in Kalman filters by introducing Kurtosis observers in the observation and state systems. The biases introduced by the chosen models will also be tracked by introducing in the state model a basis as a model of the biases (the first decomposition that will be assessed is polynomial series).

Supervision : 

Under supervision of Associate Professor  Franck RENAUD (ISAE-Supméca) and Prof Jean-Luc DION (ISAE-Supméca)