Application of network and causality based approach towards predicting onset of aeroelastic instability

We investigate the dynamical characteristics corresponding to the structural fluctuations of a cantilever suspended in a turbulent flow. First, we explore the ability of network analysis to identify the different dynamical states and probe the viability of using quantifiers of network topology as precursors for the onset of aeroelastic flutter. By increasing the flow rate or Reynolds number of the jet quasi-steadily, we observe that the structural oscillations, measured using a strain gauge, transition from low amplitude chaotic oscillations to periodic large amplitude oscillations associated with flutter. We characterize the dynamical states of the system for all these Re by constructing the weighted correlation network (CN) from the time series of strain and identifying the network properties which can be used as precursors for the onset of aeroelastic flutter. Furthermore, we illustrate the evolution of mutual statistical influence between the structural oscillations and the flow field by using Pearson correlation. We use this information in conjunction with Granger causality to identify the causal dependence between the structural oscillations and velocity fluctuations. We identify the causal variable during each dynamical regime at different regions of the flow field. Therefore, we illustrate the directional dependence through a 'cause-effect' relationship in this flow-structure interaction as it transitions to an aeroelastic flutter.