Search for: All records

Limit-cycle oscillations of bodies with airfoil cross-sections is a subject of keen interest for engineering applications. In systems consisting of multiple such closely spaced bodies, the aerodynamic interactions amongst two or more such bodies can influence the system response. The nature of these interactions is examined with respect to variations in external parameters such as freestream speed and system parameters such as inter-oscillator spacing and the number of airfoil oscillators. By using a co-simulation scheme, which consists of a reduced order three degree-of-freedom piezostructural system and an unsteady vortex lattice method fluid solver, the effects of these parameters on the resulting aerodynamic loads on the system, the overall dynamic response, and the critical flutter speed are studied. In a three-airfoil oscillator system, the effect of the position of the inner airfoil oscillator is extensively studied with a focus on characterizing airfoil interactions and airfoil-wake interactions. For different parallel configurations, studies of bifurcations with respect to different control parameters are conducted. 

With the computational resources becoming available, data-driven methods have emerged as powerful means for equation discovery and model construction. Sparse regression methods such as SINDy (Sparse Identification for Nonlinear Dynamical Systems) can be used for developing reduced-order models of nonlinear systems. In this study, the authors examine how SINDy can be used for developing low-dimensional models for airfoil systems, which experience unsteady aerodynamic loads and flutter instabilities. For a system of multiple closely spaced airfoil oscillators, analytical models are not readily available to determine flutter instabilities, and one has to take recourse to experimental and numerical means. In this work, as a starting point, data collected through simulations of unsteady aerodynamics of a single airfoil oscillator system are considered and a reduced-order model is constructed based on this data.