Convolution and Volterra Series Approach to Reduced-Order Modeling of Unsteady Aerodynamic Loads

Abstract

A combined approach of linear convolution and higher-order Volterra series (VS) to reduced-order modeling of unsteady transonic aerodynamic loads is presented. Our framework offers a simple method to determine the memory depth of the system, clearly separates the linear and the nonlinear (NL) contributions, and significantly reduces the effort required to generate a model for a wide range of reduced frequencies. The generated models use step-response data in the linear and the NL ranges, thus separating the model generation from any specific input signal or a particular reduced frequency. The models were compared against and validated using the computational fluid dynamics results of an oscillating two-dimensional NACA 0012 airfoil.