Cluster and phase-based analysis of gust-induced aeroelastic flutter

Flutter, resulting from the interaction between aerodynamic forces and structural elasticity, can cause destructive oscillations. While energy maps, constructed by studying the system's steady-state responses to sinusoidal oscillations, can offer insights into the system's response to gusts, they have limitations in capturing transient dynamics. This study develops data-driven models for transient and steady-state responses using cluster-based and phase-based strategies. We conduct numerical simulations of gust-induced aeroelastic flutter on a NACA0015 airfoil at a low Reynolds number of R_e = 1000. Data-driven phase-reduction techniques are utilized to characterize the phase sensitivity of the aeroelastic system. By examining the transient evolution of the system under gust conditions, we perform unsupervised clustering of features that capture the flutter dynamics. Key transition enablers for flutter are identified, leading to the development of cluster-based predictive models describing the system's evolution. The insights from this analysis will inform the development of control strategies to mitigate flutter oscillations.